- Email: [email protected]
A phase II trial of neoadjuvant cisplatin–fluorouracil followed by postoperative intraperitoneal floxuridine–leucovorin in patients with locally advanced gastric cancer
original article
Annals of Oncology 17: 1404–1411, 2006 doi:10.1093/annonc/mdl133 Published online 20 June 2006
A phase II trial of neoadjuvant cisplatin–fluorouracil followed by postoperative intraperitoneal floxuridine–leucovorin in patients with locally advanced gastric cancer B. Brenner, M. A. Shah, M. S. Karpeh, M. Gonen, M. F. Brennan, D. G. Coit, D. S. Klimstra, L. H. Tang & D. P. Kelsen*
Received 22 December 2005; revised 13 March 2006; accepted 5 May 2006
original article
Background: The aim of the study was to evaluate the efficacy and toxicity of neoadjuvant chemotherapy with intravenous (i.v.) cisplatin and fluorouracil (5-FU), surgery and postoperative intraperitoneal (i.p.) floxuridine (FUdR) and leucovorin (LV) in patients with locally advanced gastric cancer. Patients and methods: Preoperative staging was confirmed by laparoscopy (LAP). Two cycles of i.v. cisplatin (20 mg/m2/day, rapid infusion) and 5-FU (1000 mg/m2, continuous 24-h infusion), given on days 1–5 and 29–34, were followed by a radical gastrectomy and a D2 lymphadenectomy. Patients having R0 resections were to receive three cycles of i.p. FUdR (1000 mg/m2) and LV (240 mg/m2), given on days 1–3, 15–17 and 29–31. Intraperitoneal chemotherapy was begun 5–10 days from surgery. Results: Thirty-eight patients were treated. Both preoperative and postoperative chemotherapy were well tolerated. T stage downstaging (pretreatment LAP versus surgical pathological stage) was seen in 23% of patients. The R0 resection rate was 84%. Neither an increase in postoperative morbidity nor operative mortality was noted. With a median follow-up of 43.0 months, 15 patients (39.5%) are still alive (median survival 30.3 months). Good pathologic response, seen in five patients (15%), was associated with better survival (P = 0.053). Peritoneal and hepatic failures were found in 22% and 9% of patients, respectively. Quality of life seemed to be preserved. Conclusions: Neoadjuvant cisplatin/5-FU followed by postoperative i.p. FUdR/LV can be safely delivered to patients undergoing radical gastrectomy and D2 lymphadenectomy. The R0 resection and the survival rates are encouraging. An association between pathologic response and patient outcome was suggested. Key words: locally advanced gastric cancer, neoadjuvant chemotherapy, intraperitoneal treatment
introduction Gastric cancer is still one of the most prevalent human malignancies; 876 000 new cases were expected worldwide in the year 2000 [1]. In the United States, 22 400 new cases were expected in the year 2003 and 12 100 patients will die of the disease [2]. While the incidence of distal tumors is stable to slightly decreasing in the USA, the incidence of proximal gastric and gastroesophageal junction (GEJ) cancers, which appear to be more virulent, is rising rapidly, particularly among young Caucasian men [3]. Currently, surgery remains the mainstay of curative treatment. However, only an R0 resection is associated with *Correspondence to: Dr D. P. Kelsen, The Gastrointestinal Oncology Service, Department of Medicine, Memorial Sloan-Kettering Cancer Center, New York, New York 10021, USA. Tel: +1-212-639-8470; Fax: +1-646-422-2107; E-mail: [email protected]
ª 2006 European Society for Medical Oncology
significant cure rates. Patients having microscopic (R1) or macroscopic (R2) positive margins have an extremely poor prognosis [4]. A recent review of the USA National Cancer Database indicated that only 30% of all recently diagnosed patients underwent R0 resections [5]. Even with a potentially curative resection, 5-year survival remains less than 40%. This suggests that at time of surgery, micrometastatic distant or residual locoregional disease is already present in the majority of cases. Recently, the Intergroup-116 study demonstrated a survival advantage for the use of postoperative chemoradiation in the treatment of this disease [6]. R0 patients were randomized to adjuvant chemoradiation versus observation; improved survival (median of 36 versus 27 months, P = 0.005) was noted. Adjuvant chemoradiation, as reported by Intergroup-116, is now considered a standard care option [7–9]. In spite of the apparent
Downloaded from http://annonc.oxfordjournals.org/ at Bibliotheque de l'Universite Laval on July 12, 2015
Gastrointestinal Oncology Service, Department of Medicine, the Gastric and Mixed Tumor Service, Department of Surgery, the Department of Epidemiology and Biostatistics and the Department of Pathology, Memorial Sloan-Kettering Cancer Center, and the Weill School of Medicine, Cornell University of New York, New York 10021, USA
original article
Annals of Oncology
patients and methods patients Patients had microscopically confirmed adenocarcinoma of the stomach or GEJ. They had to have locally advanced tumors, staged (by physical examination, imaging studies and laparoscopy) as T2, N1–2M0 or T3–4NanyM0, according to the 1992 American Joint Committee on Cancer
Volume 17 | No. 9 | September 2006
criteria (AJCC) [19]. Based on data indicating that laparoscopy (LAP) was superior to endoscopic ultrasonography (EUS) in identifying high versus low risk patients and in excluding patients with unsuspected peritoneal disease (unpublished data), we did not require EUS staging as an entry criteria for this study. Later, the newer version of the AJCC, introduced in 1997 [20], was used and the results in this report are presented according to that version. Any sites of suspected M1 disease had to be ruled out prior to entrance into the study. While there were no age restrictions, a Karnofsky Performance Status (KPS) ‡60% and the ability to tolerate the proposed surgical and chemotherapy plans were necessary. No prior chemotherapy or radiation was allowed. Acceptable hematologic (WBC ‡4000 cells/mm3, platelets ‡150,000 cells/mm3), renal (BUN £30 mg/dl, creatinine £1.5 mg/dl and/or creatinine clearance >50 ml/min) and hepatic function (total serum bilirubin <2 mg/dl) was required. No clinically significant auditory impairment was allowed. Patients with prior cancer diagnosed during the previous 5-year period (except for basal cell carcinoma of skin or cervical carcinoma in situ) were excluded. Other exclusion criteria included significant cardiac disease (New York Heart Association class ‡3), pregnancy or serious intercurrent infections. The protocol was reviewed and approved by the Institutional Review Board of Memorial Sloan-Kettering Cancer Center. Written informed consent was obtained from each patient. Pre-study evaluation included a complete blood count (CBC) and differential count, a comprehensive blood profile and abdominal and pelvic computed tomography (CT). Audiogram was obtained if clinically indicated. LAP with laparoscopic ultrasound (LUS), if appropriate, was performed in all patients to exclude occult M1 disease in the peritoneum or other intra-abdominal sites. Female patients with unclear adnexal masses required gynecological examination, ultrasound and, if necessary, biopsy to exclude ovarian metastases.
treatment plan The LAP procedure involved the placement of a 10-mm infraumbilical port, with additional 5–12-mm ports as clinically indicated, for organ manipulation and specimen retrieval. The entire peritoneal surface, liver, stomach and omentum were visualized and suspicious lesions were biopsied. LUS was performed when possible, to evaluate posterior tumors better and to examine the entire liver in a systematic fashion. Peritoneal lavage was performed for cytologic analysis. Enlarged regional lymph nodes were not biopsied but were noted, if suspicious for malignancy. Suspected lymph nodes in the hepatoduodenal ligament, colonic mesentery, or the ligament of Treitz were biopsied or removed. Jejunostomy enteral nutrition tubes were not employed. Patients found to have locally advanced gastric cancers as defined above, received two cycles of cisplatin/5-FU every 28 days, followed by laparotomy and resection. Neoadjuvant chemotherapy consisted of cisplatin 20 mg/m2, as a rapid i.v. infusion daily for 5 consecutive days, and 5-FU 1000 mg/m2 as a continuous 120-h i.v. infusion (both given on days 1–5 and 29–34). Doses of both agents were attenuated for cycle 2 as appropriate for grade ‡3 toxicities; standard attenuation criteria were used. Surgery involved a radical resection, the extent (total or subtotal gastrectomy) depending on the site of the primary tumor, with a D2 lymphadenectomy. Frozen section confirmation of negative margins was recommended. In patients who had an R0 resection, an i.p. Tenkoff-type catheter was placed at the end of the procedure. Intraperitoneal chemotherapy was begun within 5 days of surgery. If postoperative complications required, therapy could be delayed for up to 10 days. Intraperitoneal therapy was not delivered if it could not be started within that time. Intraperitoneal treatment involved i.p. FUdR 1000 mg/m2 plus LV 240 mg/m2, given daily for 3 consecutive days. Three i.p. cycles, 2 weeks apart (days 1, 15, 29), were given. Intraperitoneal treatment volumes were adjusted to patient body surface area (BSA). For patients with a BSA of
doi:10.1093/annonc/mdl133 | 1405
Downloaded from http://annonc.oxfordjournals.org/ at Bibliotheque de l'Universite Laval on July 12, 2015
improvement demonstrated by Intergroup 116, half of this relatively favorable subgroup of R0 patients will still die within three years of surgery; clearly, improved therapeutic strategies are needed. Neoadjuvant (preoperative) chemotherapy represents an investigational option. Its rationale is based on the difficulty of performing an R0 resection in patients with locally advanced tumors and the high risk of micrometastatic disease in these patients. The goals of neoadjuvant treatment are to downstage the primary tumor, resulting in a higher R0 resection rate, and to simultaneously treat micrometastases early. Practically, it is employed when the patient is best able to tolerate it. This approach is supported by a recently completed large randomized study, the MAGIC trial, reported to date in abstract form. Patients with gastric and lower esophageal adenocarcinomas were assigned to pre and postoperative chemotherapy or surgery alone. Following neoadjuvant chemotherapy, significant downstaging and enhanced resectability were observed. Preliminary disease-free survival data are encouraging [10]. The peritoneal cavity is a major site for recurrent disease [11]. The use of postoperative intraperitoneal (i.p.) chemotherapy delivers chemotherapeutic agents directly into the site at high risk. Pharmacokinetic evaluation has demonstrated an advantage for i.p. over intravenous (i.v.) therapy [12]. With measurable levels in the portal vein, the liver, an equally common site of failure, also receives high doses of chemotherapy [13, 14]. Theoretically, the extremely high drug concentrations obtained with i.p. therapy may override resistance to chemotherapy. Several randomized studies have demonstrated a benefit to this approach in ovarian cancer [15]. These factors led us to perform an initial phase II trial of neoadjuvant FAMTX chemotherapy (fluorouracil, doxorubicin and methotrexate) followed by postoperative i.p. treatment with a cisplatin and fluorouracil (5FU) combination, in 56 patients with high-risk gastric cancer [16]. The resectability rate was 74% (61% R0 resections) and the median survival was 15.3 months. However, treatment was associated with significant toxicity; 60% of the patients were hospitalized at least once for neutropenic fever [16]. We performed the current phase II trial to investigate the use of a less toxic, yet active, regimen [17]. The i.p. regimen was chosen due to its minimal toxicity and the fact that floxuridine (FUdR) is almost entirely extracted on its first pass through the liver [18], making it an ideal candidate for i.p. administration [14]. The primary objectives of the trial were tolerance to the neoadjuvant regimen, its effects on operative morbidity and mortality, R0 resection rate as an immediate surrogate for eventual clinical outcome, and feasibility of postoperative i.p. FUdR/LV. Secondary objectives included downstaging, disease-free and overall survival, failure pattern and quality of life assessment.
original article £1.5, 1 l of normal saline was instilled; for patients with BSAs >1.5, 2 l were instilled on day 1 and 1 l/day on days 2–3. FUdR and LV were mixed together in the normal saline to be infused. Dose attenuation for i.p. FUdR/ LV was made for grade ‡3 mucositis, myelosuppression or diarrhea.
evaluation during the study
assessment of pathologic response Assessment of pathologic response was done independently by two pathologists (DSK, LHT), according to a method previously validated at our institution in patients with rectal and lung cancers [21, 22]. At least three microscopic sections of the tumor were examined, unless it was small enough to be submitted entirely in fewer sections. Areas of tumor treatment effect were characterized by the replacement of neoplastic glands with loosely collagenized fibrous tissue and scattered chronic inflammatory cells. Acellular mucin pools were found in cases of colloid carcinoma. Response was recorded as the histologic response percentage, i.e. the percentage of cancer cells replaced by fibrous tissue and inflammatory infiltrates. Scoring ranged from 0% (no treatment effect) to 100% (a complete response: no viable tumor identified). A score ‡60% was considered a good pathologic response.
biostatistical considerations The primary end points of the study were tolerance to neoadjuvant and postoperative chemotherapy, operative morbidity and mortality, and R0 resection rate. Secondary end points were downstaging, failure pattern, disease-free and overall survival, and quality of life. The trial was designed to accrue a total of 50 patients over 2–2.5 years. The study tested the hypothesis that the curative resection rate (R0) with neoadjuvant cisplatin– fluorouracil would be ‡60%, with a historical control at our institution of a R0 resection rate for this high-risk population of 35%–40%, with no increase in operative morbidity and mortality. Accrual to the study was lower than anticipated (38 evaluable patients accrued in 3.2 years). At that point, we determined that continuing to accrue patients would be highly unlikely to change the conclusion of this phase II trial, the purpose of which was to test whether further (definitive) study was appropriate. If 37 patients out of the original planned sample size of 50 underwent R0 resections, the resulting confidence interval would have excluded 60%. At the time we stopped accrual, 32 out of the 38 (84%) patients had undergone R0 resection. Assuming that the remaining 12 patients (if accrued) would follow the same likelihood of resection that we evaluated (using the cumulative distribution of the binomial), the probability of observing at least five R0 resections in 12 patients turned out to be 0.999. This is sufficient evidence to decide that accruing the remaining 12 patients was very unlikely to alter the conclusion. The sample analogous was used to estimate
1406 | Brenner et al.
the true R0 resection rate, operative morbidity and mortality and incidence of treatment related grade 3–4+ toxicity. Overall survival (OS) and disease-free survival (DFS) were recorded from first day of neoadjuvant treatment and from surgery, respectively. Survival distributions were estimated using the Kaplan–Meier product-limit method [23]. Comparison of survival estimates between subgroups according to pathologic response was performed using the log-rank test [24]. Quality of life was measured using FACT-G, a well validated and widely used instrument. Four subscales of FACT-G were used: physical, social and family well-being, emotional well-being and functional well-being. Each subscale’s score is adjusted to a maximum score of 100. A total score is also reported by averaging the subscores. Questionnaires were conducted at the start of neoadjuvant chemotherapy, pre-surgery and at eight time points (1, 2, 3, 6, 9, 12, 15, 18 months) after postoperative treatment. Time windows were created to group the longitudinal observations into distinct categories. Compliance was calculated as the number of questionnaires returned divided by the total number of patients alive at that time point.
results patient population Between December 1996 and February 2000, 38 patients with locally advanced gastric cancer were enrolled into the study. As shown in Table 1, the gender distribution was equal. The median age was 53 years and approximately 73% of patients were caucasian. Fifty-three per cent of patients had proximal tumors, 89% had T3–T4 tumors and 79% had N+ disease. toxicity Grade 3–4+ toxicities associated with preoperative i.v. cisplatin/ 5FU and postoperative i.p. FUdR/LV are depicted in Table 2. Treatment was generally well tolerated. The major toxicities Table 1. Patient characteristics (n = 38) Number (%) Age, years Median (range) Karnofsky performance status Median (range) Gender Male Female Ethnicity Caucasian Hispanic Asian Black Anatomic location Proximal Body Distal Pretreatment T stage T2 T3 T4 Pretreatment N status N0 N+
53 (27–77) 90 (70–90) 19 (50.0) 19 (50.0) 28 5 2 1
(73.6) (13.2) (8.8) (4.4)
20 (52.6) 8 (21.0) 10 (26.4) 4 (10.5) 32 (84.2) 2 (5.3) 8 (21.0) 30 (79.0)
Volume 17 | No. 9 | September 2006
Downloaded from http://annonc.oxfordjournals.org/ at Bibliotheque de l'Universite Laval on July 12, 2015
Physical examination and assessment of toxicity were performed prior to each cycle. Toxicity was recorded and graded according to National Cancer Institute Common Toxicity Criteria (NCI-CTC) version 1 scale. A CBC and differential count were required prior to each treatment. Comprehensive blood profiles were performed prior to each preoperative cycle. Abdominal and pelvic CT was performed after the second cycle, prior to surgery. A second LAP was repeated immediately prior to laparotomy, during the same anesthesia, for assessment of possible development of metastatic disease and of downstaging. Following completion of chemotherapy and surgery, patients were followed monthly for 3 months, then at 3 monthly intervals until year 2. At that point, 6-month follow-up visits were performed. CT scans were performed at months 3 and 6 and every 6 months thereafter. Appropriate blood studies were performed at each evaluation. A quality of life assessment using the Functional Assessment of Gastric Cancer (FACT-G) was performed prior to and periodically after treatment.
Annals of Oncology
original article
Annals of Oncology
were neutropenia and stomatitis. Twelve patients (32%) required brief (<1 week) admissions for treatment-related toxicity, usually nausea/vomiting or dehydration, two of them twice. A 2-week treatment break, during preoperative therapy, was required in one patient. There were no chemotherapyrelated deaths. Treatment was not associated with increased surgical morbidity (see below).
Table 2. Grade 3–4 toxicitiesa
Hematological Leukopenia Neutropenia Thrombocytopenia Non-hematological Stomatitis Peripheral neuropathy Diarrhea Fatigue Nausea Vomiting Dehydration
Pre-op i.v. chemotherapy Grade 3 Grade 4 (%) (%)
Post-op i.p. chemotherapy Grade 3 Grade 4 (%) (%)
8 (21%) 12 (32%) 2 (6%)
3 (9%) 14 (37%)b 0
5 (19%) 7 (27%) 1 (4%)
3 (11%) 10 (38%)2 0
5 (13%) 1 (3%)
6 (16%) 0
4 (15%) 1 (4%)
5 (19%) 0
0 8 7 2 2
1 (3%) 0 0 0 0
0 5 4 2 1
1 (4%) 0 0 0 0
(21%) (18%) (6%) (6%)
(19%) (15%) (8%) (4%)
a
Thirty-eight patients started pre-op treatment and 26 started post-op treatment. Grades are given according to the National Cancer Institute Common Toxicity Criteria (NCI-CTC) scale. b Including one episode of neutropenic fever during pre-op and one during post-op chemotherapy.
Table 3. Ability to deliver chemotherapy a
Number of patients Pre-op i.v. Cycle 1 Cycle 2 Post-op i.p. Cycle 1 Cycle 2 Cycle 3
38 32b 26 23 18
a
Dose delivered (%) Cisplatin 100 96 FUdR 96 97 98
Thirty-eight patients for the pre-op treatment and 32 having R0 resections for the post-op treatment. b See text.
Volume 17 | No. 9 | September 2006
5FU 100 95 LV 96 97 98
downstaging In the absence of a randomized phase III trial, definitive conclusions regarding the superiority of preoperative chemotherapy over surgery alone cannot be made. Downstaging of overall and T stage, among other surrogate end points, was used to evaluate potential therapeutic efficacy in this phase II trial. Pretreatment TNM stage was determined by CT and LAP, with LUS in selected patients. While a post-neoadjuvant treatment CT and LAP were obtained prior to laparotomy, for patients undergoing surgical resection, post-treatment staging was obtained at surgery and confirmed pathologically. Surgical pathological T stage was based on the deepest extent of viable tumor. Summarized in Table 4 are pretreatment laparoscopic versus post-treatment surgical pathological stages. Of the 35 operated patients for whom data regarding surgical pathologic staging was available, 10 (29%) had a decrease of at least one level in their T stage and three (9%) with N+ disease had post-treatment N0 disease. There was, however, no change in overall TNM stage, with six patients identified to have M1 disease, preoperatively or during surgery. operative outcome All patients entered into this trial had potentially resectable tumors. From our historical database of T3–T4 patients, R0 Table 4. Downstaging—pre versus post-chemotherapy
T Stage (n = 35)a T0 T1 T2 T3 T4 N status (n = 35)a N0 N+ M status (n = 38)a M0 M1 Change in overall stage Downstaged Upstaged No change
Pre-chemotherapy (laparoscopy)
Post-chemotherapy (surgical pathology)
0 0 4 30 1
0 1 12 19 3
8 27
11 24
38 0 (n = 37)b
32 6 9 11 17
a
Post-chemotherapy overall stage information is available for all 38 patients. Post-chemotherapy T and N information was missing for three patients (two who were not explored surgically and one who was operated in another institution and data regarding his post-treatment staging is incomplete). b Relevant to 37 patients; one patient was operated in another institution and data regarding his post-treatment staging is incomplete.
doi:10.1093/annonc/mdl133 | 1407
Downloaded from http://annonc.oxfordjournals.org/ at Bibliotheque de l'Universite Laval on July 12, 2015
treatment delivery The number of patients receiving pre- and postoperative chemotherapy and the percentage of planned dose delivered are summarized in Table 3. While all patients received cycle 1, six of them did not receive cycle 2 of preoperative chemotherapy, due to physician’s impression of poor tolerance (three patients), disease progression, patient refusal or emergency surgery for perforated gastric wall (one patient each). Seventy-two per cent of patients who underwent an R0 resection, and who were therefore candidates for postoperative
i.p. FUdR/LV, received at least two cycles of such treatment. In most cases when i.p. treatment could not be initiated, this was due to postoperative complications. Eighty-four per cent of patients received both preoperative cycles and 56% of those eligible received all three postoperative cycles.
original article
Annals of Oncology
(Figure 1). Of the 32 patients who had R0 resections, 12 (37.5%) are still disease-free, with a median DFS of 30.3 months (95% CI 15.2 – 43.4 months). Both the 3-year OS for the entire group and 3-year DFS for the R0 patients were 45%.
pathologic response and survival Inter-observer concordance regarding pathologic response scoring was very high; 88% of the scores were within a 10% discrepancy (ccc = 0.92) [25]. Of the 34 patients who were operated at our institution and whose specimens were available for examination by our pathologists, five good (‡60% response) pathologic responses (14.7%), none of which was complete, were observed. Pathologic response seemed to correlate with patient outcome, with borderline statistical significance; patients with good response had better survival (80% versus 31%, P = 0.053) (Figure 2). When the analysis was restricted to R0 patients (n = 30), the difference was not significant (P = 0.08).
overall and disease-free survival Median follow-up was 43.0 months, representing a relatively mature study. Fifteen patients (39.5%) are still alive, with a median OS of 30.3 months (95% CI 19.6–47.6 months) Table 5. Operative outcome
Patients eligible At risk for operation R0 resection R1 resection R2 resection No operation (M1 disease)
Number
%
38 38 32 1 3 2
100.0 84.2 2.6 7.9 5.3 Figure 1. Overall survival (n = 38).
Table 6. Postoperative complications with pre-op chemotherapy (n = 34)a
Patients with no complications Patients with complicationsb Anastomotic leak Wound infection Intra-abdominal abscess Infection Fistula Bleed Ascites leak Pneumo/hydrothorax Deep vein thrombosis Port removal
Number
%
25 10 3 2 1 1 1 1 1 2 2 1
73 29 9 6 3 3 3 3 3 6 6 3
a
Two patients were not explored due to M1 disease and two patients were operated in other institutions and data regarding their surgery is incomplete. b Five patients had two complications, five patients had one complication. The percentage of each complication relates to the number of patients out of the 35 treated patients.
1408 | Brenner et al.
Figure 2. Pathological response and survival. Data represents 34 patients who were operated at our institution and whose surgical specimens were available for examination by two independent pathologists at our institution.
Volume 17 | No. 9 | September 2006
Downloaded from http://annonc.oxfordjournals.org/ at Bibliotheque de l'Universite Laval on July 12, 2015
resections have been performed in approximately 35%–40%. We recognize that this was in a group of patients who had not had preoperative laparoscopy; recent data seem to suggest higher R0 resection rates in carefully selected patients undergoing such a procedure. Assessment of operative outcome in this study included the R0 resection rate and operative morbidity and mortality. Thirty-six patients (95%) underwent laparotomy (Table 5). Two patients developed clinically evident metastases during chemotherapy and underwent neither repeat LAP nor laparotomy. Two patients were operated in other institutions and incomplete records of their surgery were available for us. Thirty-two patients (84%) had R0 resections. Compared with patients who underwent identical surgery for gastric cancer at our institution during the same time period, but did not receive preoperative therapy, neither unexpected surgical complications nor increased rate of complications were observed (Table 6). There were no operative deaths.
Annals of Oncology
failure pattern Studies in patients undergoing surgery alone have identified the peritoneum and liver as the most common sites for recurrence after an R0 resection. These data played a major role in our decision to include postoperative i.p. therapy in the treatment program. Determined clinically (by routine imaging studies and endoscopy), 20 recurrences were observed among the 32 R0 patients. Seven peritoneal and three hepatic failures have been noted to date. Of these patients, three did not receive any i.p. therapy. Other recurrence sites were locoregional (four patients), distant nodal (two), skin (two), lung (one) and bone (one).
discussion The results of the current study indicate that neoadjuvant cisplatin/5-FU followed by postoperative i.p. FUdR/LV is safe and well tolerated by patients with locally advanced gastric cancer. There was no increase in operative morbidity and no operative mortality. Failure pattern data suggests possibly decreased peritoneal and hepatic recurrences. The R0 resection rate, in this high-risk population, is encouraging. With a fairly mature follow-up, the OS and DFS rates in patients with locally advanced gastric cancer are also encouraging. The study also suggests reproducibility and prognostic predictive value of pathologic tumor response assessment in gastric cancer. The present study demonstrated a 29% downstaging rate of T stage. Unpublished data from our institution showed an accuracy of approximately 80% between LAP and pathological T stage (data not shown). Therefore, the magnitude of ‘downstaging’ observed seems to be within the possible error of LAP proven T staging and should be viewed as such. This pattern of a decrease in T stage and only marginal effect on N stage was also noted in the MAGIC trial. In that trial, the rate of T and N pathological downstaging was 18% and 9%, respectively [26]. The i.p. component of our program was aimed at reducing the risk of peritoneal and hepatic failures. Following this approach, we observed 22% and 9% of peritoneal or liver recurrences, as the first site of failure, respectively. These rates, and especially the later, are encouraging as they seem to be on the lower range of previously reported figures, including those studies in which failure was determined clinically [6, 27, 28]. Earlier phase II studies, by our group and others, have explored a variety of neoadjuvant combinations [29–33]. These all involved chemotherapeutic regimens with modest to moderate activity against gastric cancer. For example, Siewert et al. [31] treated 41 patients using a cisplatin/5-FU/LV regimen.
Volume 17 | No. 9 | September 2006
Tolerance to chemotherapy was acceptable and R0 resection rate was 73%. We had earlier evaluated the use of preoperative FAMTX and postoperative i.p. cisplatin/5FU [32]. The R0 resection rate was 61% but myelosuppression was profound. Altogether, various phase II neoadjuvant trials, using both cisplatin-containing and non-cisplatin-containing regimens, have shown encouragingly high R0 resection rates, in the range of 60%–80%, with acceptable toxicity. Until recently, only one small randomized neoadjuvant trial has been reported [34]. A total of 107 patients were assigned to preoperative cisplatin/5-FU/etoposide or surgery only. There were no significant differences in R0 resection rates nor in survival between the two groups. Like many randomized adjuvant trials, this study was seriously statistically underpowered and a definitive statement regarding efficacy could not be made. The recently completed MAGIC trial provides a larger study regarding neoadjuvant chemotherapy in gastric cancer. In this study, 503 patients were randomized to three cycles of pre- and three cycles of postoperative epirubicin/ cisplatin/5-FU (ECF) chemotherapy or surgery alone. Neoadjuvant chemotherapy was tolerable and was completed in 88% of patients. Significant downsizing (5.0 versus 3.1 cm median tumor size, P < 0.001), downstaging (54% versus 36% T1–T2 tumors, P = 0.01) and enhanced resectability (79% versus 69%, P = 0.02) were noted. Improved progression-free survival and survival were demonstrated, with an overall 5 yearsurvival of 36% versus 23% undergoing surgery alone [10]. A recent US Intergroup Trial demonstrated that postoperative chemoradiotherapy improves survival in R0 patients: patients receiving chemoradiation had significantly better 3-year OS (50% versus 41%, P = 0.005) [6]. Toxicity, while tolerable, was substantial; 41% and 32% of patients had grade 3 or 4 toxicity, respectively, including three (1%) treatment-related deaths. This treatment is now considered a standard care option for R0 patients with locally advanced disease [7–9]. Over half the patients in Intergroup-116 had less than a D1 dissection. Data from two large randomized trials comparing D1 and D2 nodal dissections, without adjuvant treatment, suggest that outcome following D1 or D2 dissections may resemble that achieved with adjuvant chemoradiation [35, 36]. Whether or not more adequate surgery would yield a similar improvement as in Intergroup-116, it would still apply for only a minority of patients. The primary inclusion criterion in this study was an R0 resection. According to the US National Cancer Database, this is achieved in only 30% of all gastric cancer patients [5]. Understanding the factor of patient selection, the results of earlier neoadjuvant trials and ours are encouraging, as is the 45% 3-year survival we noted. Further improvements in adjuvant and neoadjuvant therapy will probably require development of more effective chemotherapeutic regimens. During the last decade, several new agents with promising activity against gastric cancer were identified. These include paclitaxel, docetaxel and irinotecan [37, 38]. These agents are now undergoing phase II and III trials, as part of combination regimens. If improved outcome is seen in advanced disease, these agents will undergo extensive testing in the adjuvant setting. Pilot trials are already under way. Response to neoadjuvant therapy has been shown, in several models (e.g. breast, lung and rectal cancers), to predict
doi:10.1093/annonc/mdl133 | 1409
Downloaded from http://annonc.oxfordjournals.org/ at Bibliotheque de l'Universite Laval on July 12, 2015
quality of life assessment The number of quality of life questionnaires completed exhibited a sharp decrease over time, with a compliance rate of 40%–70%. There was an increasing trend in the total score until period 4 (6 months), followed by a slight decrease between periods 4 and 7 (6–15 months). Understanding that the number of patients at risk was small and noting the compliance drop in later time points, no significant change in quality of life during treatment was suggested.
original article
original article
acknowledgements The authors wish to thank Aminry Reyna for her assistance in collecting the data. The study was supported in part by Grant no. RO1 CA56225 from the National Cancer Institute, Bethesda, MD.
references 1. Parkin DM, Bray FI, Devesa SS. Cancer burden in the year 2000. The global picture. Eur J Cancer 2001; 37 (Suppl 8): S4–66. 2. Jemal A, Murrey T, Samuels A et al. Cancer statistics, 2003. CA Cancer J Clin 2003; 53: 5–26. 3. Blot WJ, Devesa SS, Kneller RW, Fraumeni JF. Rising incidence of adenocarcinoma of the esophagus and gastric cardia. J Am Med Assoc 1999; 265: 1287–1289. 4. Hermanek P, Maruyama K, Sobin LH. Stomach Cancer. In Hermanek P, Gospodarowicz MK, Henson D et al. (eds): Prognostic Factors in Cancer. Berlin: Springer 1995. 5. Hundahl S, Phillips J, Menck H. The National Cancer Data Base report on poor survival of U.S. gastric carcinoma patients treated with gastrectomy: fifth edition American Joint Committee on Cancer staging, proximal disease, and the ‘different disease’ hypothesis. Cancer 2000; 88: 921–932. 6. MacDonald J, Smalley S, Benedetti J et al. Chemoradiotherapy after surgery compared with surgery alone for adenocarcinoma of the stomach or gastroesophageal junction. N Engl J Med 2001; 345: 725–730. 7. Kelsen DP. Postoperative adjuvant chemoradiation therapy for patients with resected gastric cancer: Intergroup 116. J Clin Oncol 2000; 18: 32s–34s. 8. Smalley S, Gunderson L, Tepper J et al. Gastric surgical adjuvant radiotherapy consensus report: rationale and treatment implementation. Int J Radiation Oncology Biol Phys 2002; 52: 283–293. 9. Shah MA, Kelsen DP. Post-operative adjuvant chemoradiotherapy in high risk gastric cancer. Der Chirurg 2002; 73: 325–330. 10. Cunningham D, Allum WH, Stenning SP, Weeden S. Perioperative chemotherapy in operable gastric and lower oesophageal cancer: final results of a randomised, controlled trial (the MAGIC trial, ISRCTN 93793971). J Clin Oncol 2005; 23 (16S, Part I of II): 308S. 11. Gunderson LL SH. Adenocarcinoma of the stomach: areas of failure in a reoperation series (second or symptomatic look) clinicopathologic correlation and implications for adjuvant therapy. Int J Radiation Oncol Biol Phys 1992; 8: 1–11. 12. Dedrick R. Theoretical and experimental bases of intraperitoneal chemotherapy. Semin Oncol 1985; 12 (3 Suppl 4): 1–6.
1410 | Brenner et al.
13. Speyer JL, Sugarbaker PH, Collins JM et al. Portal levels and hepatic clearance of 5-fluorouracil after intraperitoneal administration in humans. Cancer Res 1981; 41: 1916–1922. 14. Kelsen DP SL, Cohen AM, Yao TJ et al. Related articles. A phase I trial of immediate postoperative intraperitoneal floxuridine and leucovorin plus systemic 5-fluorouracil and levamisole after resection of high risk colon cancer. Cancer 1994; 74: 2224–2233. 15. Alberts DS, Markman M, Armstrong D et al. Intraperitoneal therapy for stage III ovarian cancer: a therapy whose time has come! J Clin Oncol 2002; 20: 3944–3946. 16. Kelsen DP. Adjuvant and neoadjuvant therapy for gastric cancer. Semin Oncol 1996; 23: 379–389. 17. Vanhoefer U, Rougier P, Wilke H et al. Final results of a randomized phase III trial of sequential high-dose methotrexate, fluorouracil, and doxorubicin versus etoposide, leucovorin, and fluorouracil versus infusional fluorouracil and cisplatin in advanced gastric cancer: a trial of the European Organization for Research and Treatment of Cancer Gastrointestinal Tract Cancer Cooperative Group. J Clin Oncol 2000; 18: 2648–2657. 18. Ensminger WD, Rosowsky A, Raso V et al. A clinical-pharmacological evaluation of hepatic arterial infusions of 5-fluoro-29-deoxyuridine and 5-fluorouracil. Cancer Res 1978; 38 (11 part 1): 3784–3792. 19. Beahrs OH HD, Hutter RVP, Kennedy BJ (eds). Manual for Staging of Cancer. American Joint Committee on Cancer. Philadelphia: J.B. Lippincott 1992. 20. Fleming I, Cooper J, Henson D et al. American Joint Committee on Cancer Staging Manual. Philadelphia: Lippincott-Raven 1997. 21. Ruo L, Tickoo S, Klimstra DS et al. Long-term prognostic significance of extent of rectal cancer response to preoperative radiation and chemotherapy. Ann Surg 2002; 236: 75–81. 22. Rusch V, Klimstra D, Venkatraman E et al. Aberrant p53 expression predicts clinical resistance to cisplatin-based chemotherapy in locally advanced non-small cell lung cancer. Cancer Res 1995; 55: 5038–5042. 23. Kaplan EL, Meier P. Non-parametric estimation from incomplete observation. J Am Stat 1958; 53: 457–481. 24. Mantel N. Evaluation of survival data and two new rank order statistics arising in its consideration. Cancer Chemother Rep 1966; 50: 163–170. 25. Lin LI. A concordance correlation coefficient to evaluate reproducibility. Biometrics 1989; 45: 255–268. 26. Allum W, Cunningham D, Weeden S et al. Perioperative chemotherapy and operable gastric and lower esophageal cancer: a randomized, controlled trial (the MAGIC trial ISRCNT93793971). Am Soc Clin Oncol 2003; 22: 249. 27. Estape J, Grau JJ, Lcobendas F et al. Mitomycin C as an adjuvant treatment to resected gastric cancer. A 10-year follow-up. Ann Surg 1991; 213: 219– 221. 28. Laundry J, Tepper J, Wood W et al. Patterns of failure following curative resection of gastric carcinoma. Int J Radiation Oncol Biol Phys 1990; 19: 1357– 1362. 29. Lowy AM, Mansfield PF, Leach SD et al. Response to neoadjuvant chemotherapy best predicts survival after curative resection of gastric cancer. Ann Surg 1999; 229: 303–308. 30. Crookes P, Leichman CG, Leichman L et al. Systemic chemotherapy for gastric carcinoma followed by postoperative intraperitoneal therapy: a final report. Cancer 1997; 79: 1767–1775. 31. Siewert JR, Bottcher K, Roder JD et al. Prognostic relevance of systemic lymph node dissection in gastric carcinoma. Br J Surg 1993; 80: 1015–1018. 32. Kelsen D, Karpeh M, Schwartz G, et al. Neoadjuvant therapy of high-risk gastric cancer: a phase II trial of preoperative FAMTX and postoperative intraperitoneal fluorouracil-cisplatin plus intravenous fluorouracil. J Clin Oncol 1996; 14: 1818. 33. Alexander HR, Grem JL, Hamilton JM et al. Thymidylate synthase protein expression association with response to neoadjuvant chemotherapy and resection for locally advanced gastric and gastroesophageal adenocarcinoma. Cancer J Sci Am 1995; 1: 49. 34. Kang YK, Choi DW, Im YH et al. A phase III randomized comparison of neoadjuvant chemotherapy followed by surgery versus surgery for locally advanced stomach cancer. Proc ASCO 1996; 15: 215.
Volume 17 | No. 9 | September 2006
Downloaded from http://annonc.oxfordjournals.org/ at Bibliotheque de l'Universite Laval on July 12, 2015
improved patient outcome [21, 39, 40]. Mandard et al. [41] were the first to describe a method for pathologic response assessment in upper gastrointestinal malignancies which successfully predicted outcome. In their study, on esophageal cancer patients receiving preoperative chemoradiation, pathologic response strongly correlated with DFS. A comparable method was recently evaluated in gastric cancer [29]. Using a similar approach, our results suggest that pathologic response is associated with improved outcome in gastric cancer. If confirmed in larger trials, pathologic response would allow rapid evaluation of efficacy for future neoadjuvant studies. In summary, neoadjuvant cisplatin/5-FU and postoperative i.p. FUdR/LV can be safely delivered to patients undergoing radical gastrectomy and extensive lymph node dissection. The R0 resection rate is high, with no increase in operative morbidity and mortality. The failure pattern is consistent with possibly lower rate of i.p. and hepatic failures. New trials, using the same approach and employing possibly more effective chemotherapy regimens, are underway.
Annals of Oncology
Annals of Oncology
35. Bonenkamp JJ, Songun J, Hermans J et al. Randomised comparison of morbidity after D1 and D2 dissection for gastric cancer in 996 Dutch patients. Lancet 1995; 345: 745–748. 36. Cuschieri A, Weeden S, Fielding J et al. Patient survival after D1 and D2 resections for gastric cancer: long-term results of the MRC randomized surgical trial. Surgical Cooperative group. Br J Cancer 1999; 79: 1522–1530. 37. Shah MA. Recent developments in the treatment of gastric carcinoma. Curr Oncol Rep 2002; 4: 213–221. 38. Janunger KG, Hafstrom L, Nygren P, Glimelius B. A systemic overview of chemotherapy effects in gastric cancer. Acta Onc 2001; 40: 309–326.
original article 39. Bonadonna G, Valagussa P, Brambilla C et al. Primary chemotherapy in operable breast cancer: eight-year experience at the Milan Cancer Institute. J Clin Oncol 1998; 16: 93–100. 40. Pisters KM, Kris MG, Gralla RJ et al. Pathologic complete response in advanced non-small-cell lung cancer following preoperative chemotherapy: implications for the design of future non-small-cell lung cancer combined modality trials. J Clin Oncol 1993; 11: 1757–1762. 41. Mandard AM, Dalibard F, Mandard JC et al. Pathologic assessment of tumor regression after preoperative chemoradiotherapy of esophageal carcinoma: clinicopathologic correlations. Cancer 1994; 73: 2680–2686.
Downloaded from http://annonc.oxfordjournals.org/ at Bibliotheque de l'Universite Laval on July 12, 2015
Volume 17 | No. 9 | September 2006
doi:10.1093/annonc/mdl133 | 1411
Annals of Oncology 17: 1404–1411, 2006 doi:10.1093/annonc/mdl133 Published online 20 June 2006
A phase II trial of neoadjuvant cisplatin–fluorouracil followed by postoperative intraperitoneal floxuridine–leucovorin in patients with locally advanced gastric cancer B. Brenner, M. A. Shah, M. S. Karpeh, M. Gonen, M. F. Brennan, D. G. Coit, D. S. Klimstra, L. H. Tang & D. P. Kelsen*
Received 22 December 2005; revised 13 March 2006; accepted 5 May 2006
original article
Background: The aim of the study was to evaluate the efficacy and toxicity of neoadjuvant chemotherapy with intravenous (i.v.) cisplatin and fluorouracil (5-FU), surgery and postoperative intraperitoneal (i.p.) floxuridine (FUdR) and leucovorin (LV) in patients with locally advanced gastric cancer. Patients and methods: Preoperative staging was confirmed by laparoscopy (LAP). Two cycles of i.v. cisplatin (20 mg/m2/day, rapid infusion) and 5-FU (1000 mg/m2, continuous 24-h infusion), given on days 1–5 and 29–34, were followed by a radical gastrectomy and a D2 lymphadenectomy. Patients having R0 resections were to receive three cycles of i.p. FUdR (1000 mg/m2) and LV (240 mg/m2), given on days 1–3, 15–17 and 29–31. Intraperitoneal chemotherapy was begun 5–10 days from surgery. Results: Thirty-eight patients were treated. Both preoperative and postoperative chemotherapy were well tolerated. T stage downstaging (pretreatment LAP versus surgical pathological stage) was seen in 23% of patients. The R0 resection rate was 84%. Neither an increase in postoperative morbidity nor operative mortality was noted. With a median follow-up of 43.0 months, 15 patients (39.5%) are still alive (median survival 30.3 months). Good pathologic response, seen in five patients (15%), was associated with better survival (P = 0.053). Peritoneal and hepatic failures were found in 22% and 9% of patients, respectively. Quality of life seemed to be preserved. Conclusions: Neoadjuvant cisplatin/5-FU followed by postoperative i.p. FUdR/LV can be safely delivered to patients undergoing radical gastrectomy and D2 lymphadenectomy. The R0 resection and the survival rates are encouraging. An association between pathologic response and patient outcome was suggested. Key words: locally advanced gastric cancer, neoadjuvant chemotherapy, intraperitoneal treatment
introduction Gastric cancer is still one of the most prevalent human malignancies; 876 000 new cases were expected worldwide in the year 2000 [1]. In the United States, 22 400 new cases were expected in the year 2003 and 12 100 patients will die of the disease [2]. While the incidence of distal tumors is stable to slightly decreasing in the USA, the incidence of proximal gastric and gastroesophageal junction (GEJ) cancers, which appear to be more virulent, is rising rapidly, particularly among young Caucasian men [3]. Currently, surgery remains the mainstay of curative treatment. However, only an R0 resection is associated with *Correspondence to: Dr D. P. Kelsen, The Gastrointestinal Oncology Service, Department of Medicine, Memorial Sloan-Kettering Cancer Center, New York, New York 10021, USA. Tel: +1-212-639-8470; Fax: +1-646-422-2107; E-mail: [email protected]
ª 2006 European Society for Medical Oncology
significant cure rates. Patients having microscopic (R1) or macroscopic (R2) positive margins have an extremely poor prognosis [4]. A recent review of the USA National Cancer Database indicated that only 30% of all recently diagnosed patients underwent R0 resections [5]. Even with a potentially curative resection, 5-year survival remains less than 40%. This suggests that at time of surgery, micrometastatic distant or residual locoregional disease is already present in the majority of cases. Recently, the Intergroup-116 study demonstrated a survival advantage for the use of postoperative chemoradiation in the treatment of this disease [6]. R0 patients were randomized to adjuvant chemoradiation versus observation; improved survival (median of 36 versus 27 months, P = 0.005) was noted. Adjuvant chemoradiation, as reported by Intergroup-116, is now considered a standard care option [7–9]. In spite of the apparent
Downloaded from http://annonc.oxfordjournals.org/ at Bibliotheque de l'Universite Laval on July 12, 2015
Gastrointestinal Oncology Service, Department of Medicine, the Gastric and Mixed Tumor Service, Department of Surgery, the Department of Epidemiology and Biostatistics and the Department of Pathology, Memorial Sloan-Kettering Cancer Center, and the Weill School of Medicine, Cornell University of New York, New York 10021, USA
original article
Annals of Oncology
patients and methods patients Patients had microscopically confirmed adenocarcinoma of the stomach or GEJ. They had to have locally advanced tumors, staged (by physical examination, imaging studies and laparoscopy) as T2, N1–2M0 or T3–4NanyM0, according to the 1992 American Joint Committee on Cancer
Volume 17 | No. 9 | September 2006
criteria (AJCC) [19]. Based on data indicating that laparoscopy (LAP) was superior to endoscopic ultrasonography (EUS) in identifying high versus low risk patients and in excluding patients with unsuspected peritoneal disease (unpublished data), we did not require EUS staging as an entry criteria for this study. Later, the newer version of the AJCC, introduced in 1997 [20], was used and the results in this report are presented according to that version. Any sites of suspected M1 disease had to be ruled out prior to entrance into the study. While there were no age restrictions, a Karnofsky Performance Status (KPS) ‡60% and the ability to tolerate the proposed surgical and chemotherapy plans were necessary. No prior chemotherapy or radiation was allowed. Acceptable hematologic (WBC ‡4000 cells/mm3, platelets ‡150,000 cells/mm3), renal (BUN £30 mg/dl, creatinine £1.5 mg/dl and/or creatinine clearance >50 ml/min) and hepatic function (total serum bilirubin <2 mg/dl) was required. No clinically significant auditory impairment was allowed. Patients with prior cancer diagnosed during the previous 5-year period (except for basal cell carcinoma of skin or cervical carcinoma in situ) were excluded. Other exclusion criteria included significant cardiac disease (New York Heart Association class ‡3), pregnancy or serious intercurrent infections. The protocol was reviewed and approved by the Institutional Review Board of Memorial Sloan-Kettering Cancer Center. Written informed consent was obtained from each patient. Pre-study evaluation included a complete blood count (CBC) and differential count, a comprehensive blood profile and abdominal and pelvic computed tomography (CT). Audiogram was obtained if clinically indicated. LAP with laparoscopic ultrasound (LUS), if appropriate, was performed in all patients to exclude occult M1 disease in the peritoneum or other intra-abdominal sites. Female patients with unclear adnexal masses required gynecological examination, ultrasound and, if necessary, biopsy to exclude ovarian metastases.
treatment plan The LAP procedure involved the placement of a 10-mm infraumbilical port, with additional 5–12-mm ports as clinically indicated, for organ manipulation and specimen retrieval. The entire peritoneal surface, liver, stomach and omentum were visualized and suspicious lesions were biopsied. LUS was performed when possible, to evaluate posterior tumors better and to examine the entire liver in a systematic fashion. Peritoneal lavage was performed for cytologic analysis. Enlarged regional lymph nodes were not biopsied but were noted, if suspicious for malignancy. Suspected lymph nodes in the hepatoduodenal ligament, colonic mesentery, or the ligament of Treitz were biopsied or removed. Jejunostomy enteral nutrition tubes were not employed. Patients found to have locally advanced gastric cancers as defined above, received two cycles of cisplatin/5-FU every 28 days, followed by laparotomy and resection. Neoadjuvant chemotherapy consisted of cisplatin 20 mg/m2, as a rapid i.v. infusion daily for 5 consecutive days, and 5-FU 1000 mg/m2 as a continuous 120-h i.v. infusion (both given on days 1–5 and 29–34). Doses of both agents were attenuated for cycle 2 as appropriate for grade ‡3 toxicities; standard attenuation criteria were used. Surgery involved a radical resection, the extent (total or subtotal gastrectomy) depending on the site of the primary tumor, with a D2 lymphadenectomy. Frozen section confirmation of negative margins was recommended. In patients who had an R0 resection, an i.p. Tenkoff-type catheter was placed at the end of the procedure. Intraperitoneal chemotherapy was begun within 5 days of surgery. If postoperative complications required, therapy could be delayed for up to 10 days. Intraperitoneal therapy was not delivered if it could not be started within that time. Intraperitoneal treatment involved i.p. FUdR 1000 mg/m2 plus LV 240 mg/m2, given daily for 3 consecutive days. Three i.p. cycles, 2 weeks apart (days 1, 15, 29), were given. Intraperitoneal treatment volumes were adjusted to patient body surface area (BSA). For patients with a BSA of
doi:10.1093/annonc/mdl133 | 1405
Downloaded from http://annonc.oxfordjournals.org/ at Bibliotheque de l'Universite Laval on July 12, 2015
improvement demonstrated by Intergroup 116, half of this relatively favorable subgroup of R0 patients will still die within three years of surgery; clearly, improved therapeutic strategies are needed. Neoadjuvant (preoperative) chemotherapy represents an investigational option. Its rationale is based on the difficulty of performing an R0 resection in patients with locally advanced tumors and the high risk of micrometastatic disease in these patients. The goals of neoadjuvant treatment are to downstage the primary tumor, resulting in a higher R0 resection rate, and to simultaneously treat micrometastases early. Practically, it is employed when the patient is best able to tolerate it. This approach is supported by a recently completed large randomized study, the MAGIC trial, reported to date in abstract form. Patients with gastric and lower esophageal adenocarcinomas were assigned to pre and postoperative chemotherapy or surgery alone. Following neoadjuvant chemotherapy, significant downstaging and enhanced resectability were observed. Preliminary disease-free survival data are encouraging [10]. The peritoneal cavity is a major site for recurrent disease [11]. The use of postoperative intraperitoneal (i.p.) chemotherapy delivers chemotherapeutic agents directly into the site at high risk. Pharmacokinetic evaluation has demonstrated an advantage for i.p. over intravenous (i.v.) therapy [12]. With measurable levels in the portal vein, the liver, an equally common site of failure, also receives high doses of chemotherapy [13, 14]. Theoretically, the extremely high drug concentrations obtained with i.p. therapy may override resistance to chemotherapy. Several randomized studies have demonstrated a benefit to this approach in ovarian cancer [15]. These factors led us to perform an initial phase II trial of neoadjuvant FAMTX chemotherapy (fluorouracil, doxorubicin and methotrexate) followed by postoperative i.p. treatment with a cisplatin and fluorouracil (5FU) combination, in 56 patients with high-risk gastric cancer [16]. The resectability rate was 74% (61% R0 resections) and the median survival was 15.3 months. However, treatment was associated with significant toxicity; 60% of the patients were hospitalized at least once for neutropenic fever [16]. We performed the current phase II trial to investigate the use of a less toxic, yet active, regimen [17]. The i.p. regimen was chosen due to its minimal toxicity and the fact that floxuridine (FUdR) is almost entirely extracted on its first pass through the liver [18], making it an ideal candidate for i.p. administration [14]. The primary objectives of the trial were tolerance to the neoadjuvant regimen, its effects on operative morbidity and mortality, R0 resection rate as an immediate surrogate for eventual clinical outcome, and feasibility of postoperative i.p. FUdR/LV. Secondary objectives included downstaging, disease-free and overall survival, failure pattern and quality of life assessment.
original article £1.5, 1 l of normal saline was instilled; for patients with BSAs >1.5, 2 l were instilled on day 1 and 1 l/day on days 2–3. FUdR and LV were mixed together in the normal saline to be infused. Dose attenuation for i.p. FUdR/ LV was made for grade ‡3 mucositis, myelosuppression or diarrhea.
evaluation during the study
assessment of pathologic response Assessment of pathologic response was done independently by two pathologists (DSK, LHT), according to a method previously validated at our institution in patients with rectal and lung cancers [21, 22]. At least three microscopic sections of the tumor were examined, unless it was small enough to be submitted entirely in fewer sections. Areas of tumor treatment effect were characterized by the replacement of neoplastic glands with loosely collagenized fibrous tissue and scattered chronic inflammatory cells. Acellular mucin pools were found in cases of colloid carcinoma. Response was recorded as the histologic response percentage, i.e. the percentage of cancer cells replaced by fibrous tissue and inflammatory infiltrates. Scoring ranged from 0% (no treatment effect) to 100% (a complete response: no viable tumor identified). A score ‡60% was considered a good pathologic response.
biostatistical considerations The primary end points of the study were tolerance to neoadjuvant and postoperative chemotherapy, operative morbidity and mortality, and R0 resection rate. Secondary end points were downstaging, failure pattern, disease-free and overall survival, and quality of life. The trial was designed to accrue a total of 50 patients over 2–2.5 years. The study tested the hypothesis that the curative resection rate (R0) with neoadjuvant cisplatin– fluorouracil would be ‡60%, with a historical control at our institution of a R0 resection rate for this high-risk population of 35%–40%, with no increase in operative morbidity and mortality. Accrual to the study was lower than anticipated (38 evaluable patients accrued in 3.2 years). At that point, we determined that continuing to accrue patients would be highly unlikely to change the conclusion of this phase II trial, the purpose of which was to test whether further (definitive) study was appropriate. If 37 patients out of the original planned sample size of 50 underwent R0 resections, the resulting confidence interval would have excluded 60%. At the time we stopped accrual, 32 out of the 38 (84%) patients had undergone R0 resection. Assuming that the remaining 12 patients (if accrued) would follow the same likelihood of resection that we evaluated (using the cumulative distribution of the binomial), the probability of observing at least five R0 resections in 12 patients turned out to be 0.999. This is sufficient evidence to decide that accruing the remaining 12 patients was very unlikely to alter the conclusion. The sample analogous was used to estimate
1406 | Brenner et al.
the true R0 resection rate, operative morbidity and mortality and incidence of treatment related grade 3–4+ toxicity. Overall survival (OS) and disease-free survival (DFS) were recorded from first day of neoadjuvant treatment and from surgery, respectively. Survival distributions were estimated using the Kaplan–Meier product-limit method [23]. Comparison of survival estimates between subgroups according to pathologic response was performed using the log-rank test [24]. Quality of life was measured using FACT-G, a well validated and widely used instrument. Four subscales of FACT-G were used: physical, social and family well-being, emotional well-being and functional well-being. Each subscale’s score is adjusted to a maximum score of 100. A total score is also reported by averaging the subscores. Questionnaires were conducted at the start of neoadjuvant chemotherapy, pre-surgery and at eight time points (1, 2, 3, 6, 9, 12, 15, 18 months) after postoperative treatment. Time windows were created to group the longitudinal observations into distinct categories. Compliance was calculated as the number of questionnaires returned divided by the total number of patients alive at that time point.
results patient population Between December 1996 and February 2000, 38 patients with locally advanced gastric cancer were enrolled into the study. As shown in Table 1, the gender distribution was equal. The median age was 53 years and approximately 73% of patients were caucasian. Fifty-three per cent of patients had proximal tumors, 89% had T3–T4 tumors and 79% had N+ disease. toxicity Grade 3–4+ toxicities associated with preoperative i.v. cisplatin/ 5FU and postoperative i.p. FUdR/LV are depicted in Table 2. Treatment was generally well tolerated. The major toxicities Table 1. Patient characteristics (n = 38) Number (%) Age, years Median (range) Karnofsky performance status Median (range) Gender Male Female Ethnicity Caucasian Hispanic Asian Black Anatomic location Proximal Body Distal Pretreatment T stage T2 T3 T4 Pretreatment N status N0 N+
53 (27–77) 90 (70–90) 19 (50.0) 19 (50.0) 28 5 2 1
(73.6) (13.2) (8.8) (4.4)
20 (52.6) 8 (21.0) 10 (26.4) 4 (10.5) 32 (84.2) 2 (5.3) 8 (21.0) 30 (79.0)
Volume 17 | No. 9 | September 2006
Downloaded from http://annonc.oxfordjournals.org/ at Bibliotheque de l'Universite Laval on July 12, 2015
Physical examination and assessment of toxicity were performed prior to each cycle. Toxicity was recorded and graded according to National Cancer Institute Common Toxicity Criteria (NCI-CTC) version 1 scale. A CBC and differential count were required prior to each treatment. Comprehensive blood profiles were performed prior to each preoperative cycle. Abdominal and pelvic CT was performed after the second cycle, prior to surgery. A second LAP was repeated immediately prior to laparotomy, during the same anesthesia, for assessment of possible development of metastatic disease and of downstaging. Following completion of chemotherapy and surgery, patients were followed monthly for 3 months, then at 3 monthly intervals until year 2. At that point, 6-month follow-up visits were performed. CT scans were performed at months 3 and 6 and every 6 months thereafter. Appropriate blood studies were performed at each evaluation. A quality of life assessment using the Functional Assessment of Gastric Cancer (FACT-G) was performed prior to and periodically after treatment.
Annals of Oncology
original article
Annals of Oncology
were neutropenia and stomatitis. Twelve patients (32%) required brief (<1 week) admissions for treatment-related toxicity, usually nausea/vomiting or dehydration, two of them twice. A 2-week treatment break, during preoperative therapy, was required in one patient. There were no chemotherapyrelated deaths. Treatment was not associated with increased surgical morbidity (see below).
Table 2. Grade 3–4 toxicitiesa
Hematological Leukopenia Neutropenia Thrombocytopenia Non-hematological Stomatitis Peripheral neuropathy Diarrhea Fatigue Nausea Vomiting Dehydration
Pre-op i.v. chemotherapy Grade 3 Grade 4 (%) (%)
Post-op i.p. chemotherapy Grade 3 Grade 4 (%) (%)
8 (21%) 12 (32%) 2 (6%)
3 (9%) 14 (37%)b 0
5 (19%) 7 (27%) 1 (4%)
3 (11%) 10 (38%)2 0
5 (13%) 1 (3%)
6 (16%) 0
4 (15%) 1 (4%)
5 (19%) 0
0 8 7 2 2
1 (3%) 0 0 0 0
0 5 4 2 1
1 (4%) 0 0 0 0
(21%) (18%) (6%) (6%)
(19%) (15%) (8%) (4%)
a
Thirty-eight patients started pre-op treatment and 26 started post-op treatment. Grades are given according to the National Cancer Institute Common Toxicity Criteria (NCI-CTC) scale. b Including one episode of neutropenic fever during pre-op and one during post-op chemotherapy.
Table 3. Ability to deliver chemotherapy a
Number of patients Pre-op i.v. Cycle 1 Cycle 2 Post-op i.p. Cycle 1 Cycle 2 Cycle 3
38 32b 26 23 18
a
Dose delivered (%) Cisplatin 100 96 FUdR 96 97 98
Thirty-eight patients for the pre-op treatment and 32 having R0 resections for the post-op treatment. b See text.
Volume 17 | No. 9 | September 2006
5FU 100 95 LV 96 97 98
downstaging In the absence of a randomized phase III trial, definitive conclusions regarding the superiority of preoperative chemotherapy over surgery alone cannot be made. Downstaging of overall and T stage, among other surrogate end points, was used to evaluate potential therapeutic efficacy in this phase II trial. Pretreatment TNM stage was determined by CT and LAP, with LUS in selected patients. While a post-neoadjuvant treatment CT and LAP were obtained prior to laparotomy, for patients undergoing surgical resection, post-treatment staging was obtained at surgery and confirmed pathologically. Surgical pathological T stage was based on the deepest extent of viable tumor. Summarized in Table 4 are pretreatment laparoscopic versus post-treatment surgical pathological stages. Of the 35 operated patients for whom data regarding surgical pathologic staging was available, 10 (29%) had a decrease of at least one level in their T stage and three (9%) with N+ disease had post-treatment N0 disease. There was, however, no change in overall TNM stage, with six patients identified to have M1 disease, preoperatively or during surgery. operative outcome All patients entered into this trial had potentially resectable tumors. From our historical database of T3–T4 patients, R0 Table 4. Downstaging—pre versus post-chemotherapy
T Stage (n = 35)a T0 T1 T2 T3 T4 N status (n = 35)a N0 N+ M status (n = 38)a M0 M1 Change in overall stage Downstaged Upstaged No change
Pre-chemotherapy (laparoscopy)
Post-chemotherapy (surgical pathology)
0 0 4 30 1
0 1 12 19 3
8 27
11 24
38 0 (n = 37)b
32 6 9 11 17
a
Post-chemotherapy overall stage information is available for all 38 patients. Post-chemotherapy T and N information was missing for three patients (two who were not explored surgically and one who was operated in another institution and data regarding his post-treatment staging is incomplete). b Relevant to 37 patients; one patient was operated in another institution and data regarding his post-treatment staging is incomplete.
doi:10.1093/annonc/mdl133 | 1407
Downloaded from http://annonc.oxfordjournals.org/ at Bibliotheque de l'Universite Laval on July 12, 2015
treatment delivery The number of patients receiving pre- and postoperative chemotherapy and the percentage of planned dose delivered are summarized in Table 3. While all patients received cycle 1, six of them did not receive cycle 2 of preoperative chemotherapy, due to physician’s impression of poor tolerance (three patients), disease progression, patient refusal or emergency surgery for perforated gastric wall (one patient each). Seventy-two per cent of patients who underwent an R0 resection, and who were therefore candidates for postoperative
i.p. FUdR/LV, received at least two cycles of such treatment. In most cases when i.p. treatment could not be initiated, this was due to postoperative complications. Eighty-four per cent of patients received both preoperative cycles and 56% of those eligible received all three postoperative cycles.
original article
Annals of Oncology
(Figure 1). Of the 32 patients who had R0 resections, 12 (37.5%) are still disease-free, with a median DFS of 30.3 months (95% CI 15.2 – 43.4 months). Both the 3-year OS for the entire group and 3-year DFS for the R0 patients were 45%.
pathologic response and survival Inter-observer concordance regarding pathologic response scoring was very high; 88% of the scores were within a 10% discrepancy (ccc = 0.92) [25]. Of the 34 patients who were operated at our institution and whose specimens were available for examination by our pathologists, five good (‡60% response) pathologic responses (14.7%), none of which was complete, were observed. Pathologic response seemed to correlate with patient outcome, with borderline statistical significance; patients with good response had better survival (80% versus 31%, P = 0.053) (Figure 2). When the analysis was restricted to R0 patients (n = 30), the difference was not significant (P = 0.08).
overall and disease-free survival Median follow-up was 43.0 months, representing a relatively mature study. Fifteen patients (39.5%) are still alive, with a median OS of 30.3 months (95% CI 19.6–47.6 months) Table 5. Operative outcome
Patients eligible At risk for operation R0 resection R1 resection R2 resection No operation (M1 disease)
Number
%
38 38 32 1 3 2
100.0 84.2 2.6 7.9 5.3 Figure 1. Overall survival (n = 38).
Table 6. Postoperative complications with pre-op chemotherapy (n = 34)a
Patients with no complications Patients with complicationsb Anastomotic leak Wound infection Intra-abdominal abscess Infection Fistula Bleed Ascites leak Pneumo/hydrothorax Deep vein thrombosis Port removal
Number
%
25 10 3 2 1 1 1 1 1 2 2 1
73 29 9 6 3 3 3 3 3 6 6 3
a
Two patients were not explored due to M1 disease and two patients were operated in other institutions and data regarding their surgery is incomplete. b Five patients had two complications, five patients had one complication. The percentage of each complication relates to the number of patients out of the 35 treated patients.
1408 | Brenner et al.
Figure 2. Pathological response and survival. Data represents 34 patients who were operated at our institution and whose surgical specimens were available for examination by two independent pathologists at our institution.
Volume 17 | No. 9 | September 2006
Downloaded from http://annonc.oxfordjournals.org/ at Bibliotheque de l'Universite Laval on July 12, 2015
resections have been performed in approximately 35%–40%. We recognize that this was in a group of patients who had not had preoperative laparoscopy; recent data seem to suggest higher R0 resection rates in carefully selected patients undergoing such a procedure. Assessment of operative outcome in this study included the R0 resection rate and operative morbidity and mortality. Thirty-six patients (95%) underwent laparotomy (Table 5). Two patients developed clinically evident metastases during chemotherapy and underwent neither repeat LAP nor laparotomy. Two patients were operated in other institutions and incomplete records of their surgery were available for us. Thirty-two patients (84%) had R0 resections. Compared with patients who underwent identical surgery for gastric cancer at our institution during the same time period, but did not receive preoperative therapy, neither unexpected surgical complications nor increased rate of complications were observed (Table 6). There were no operative deaths.
Annals of Oncology
failure pattern Studies in patients undergoing surgery alone have identified the peritoneum and liver as the most common sites for recurrence after an R0 resection. These data played a major role in our decision to include postoperative i.p. therapy in the treatment program. Determined clinically (by routine imaging studies and endoscopy), 20 recurrences were observed among the 32 R0 patients. Seven peritoneal and three hepatic failures have been noted to date. Of these patients, three did not receive any i.p. therapy. Other recurrence sites were locoregional (four patients), distant nodal (two), skin (two), lung (one) and bone (one).
discussion The results of the current study indicate that neoadjuvant cisplatin/5-FU followed by postoperative i.p. FUdR/LV is safe and well tolerated by patients with locally advanced gastric cancer. There was no increase in operative morbidity and no operative mortality. Failure pattern data suggests possibly decreased peritoneal and hepatic recurrences. The R0 resection rate, in this high-risk population, is encouraging. With a fairly mature follow-up, the OS and DFS rates in patients with locally advanced gastric cancer are also encouraging. The study also suggests reproducibility and prognostic predictive value of pathologic tumor response assessment in gastric cancer. The present study demonstrated a 29% downstaging rate of T stage. Unpublished data from our institution showed an accuracy of approximately 80% between LAP and pathological T stage (data not shown). Therefore, the magnitude of ‘downstaging’ observed seems to be within the possible error of LAP proven T staging and should be viewed as such. This pattern of a decrease in T stage and only marginal effect on N stage was also noted in the MAGIC trial. In that trial, the rate of T and N pathological downstaging was 18% and 9%, respectively [26]. The i.p. component of our program was aimed at reducing the risk of peritoneal and hepatic failures. Following this approach, we observed 22% and 9% of peritoneal or liver recurrences, as the first site of failure, respectively. These rates, and especially the later, are encouraging as they seem to be on the lower range of previously reported figures, including those studies in which failure was determined clinically [6, 27, 28]. Earlier phase II studies, by our group and others, have explored a variety of neoadjuvant combinations [29–33]. These all involved chemotherapeutic regimens with modest to moderate activity against gastric cancer. For example, Siewert et al. [31] treated 41 patients using a cisplatin/5-FU/LV regimen.
Volume 17 | No. 9 | September 2006
Tolerance to chemotherapy was acceptable and R0 resection rate was 73%. We had earlier evaluated the use of preoperative FAMTX and postoperative i.p. cisplatin/5FU [32]. The R0 resection rate was 61% but myelosuppression was profound. Altogether, various phase II neoadjuvant trials, using both cisplatin-containing and non-cisplatin-containing regimens, have shown encouragingly high R0 resection rates, in the range of 60%–80%, with acceptable toxicity. Until recently, only one small randomized neoadjuvant trial has been reported [34]. A total of 107 patients were assigned to preoperative cisplatin/5-FU/etoposide or surgery only. There were no significant differences in R0 resection rates nor in survival between the two groups. Like many randomized adjuvant trials, this study was seriously statistically underpowered and a definitive statement regarding efficacy could not be made. The recently completed MAGIC trial provides a larger study regarding neoadjuvant chemotherapy in gastric cancer. In this study, 503 patients were randomized to three cycles of pre- and three cycles of postoperative epirubicin/ cisplatin/5-FU (ECF) chemotherapy or surgery alone. Neoadjuvant chemotherapy was tolerable and was completed in 88% of patients. Significant downsizing (5.0 versus 3.1 cm median tumor size, P < 0.001), downstaging (54% versus 36% T1–T2 tumors, P = 0.01) and enhanced resectability (79% versus 69%, P = 0.02) were noted. Improved progression-free survival and survival were demonstrated, with an overall 5 yearsurvival of 36% versus 23% undergoing surgery alone [10]. A recent US Intergroup Trial demonstrated that postoperative chemoradiotherapy improves survival in R0 patients: patients receiving chemoradiation had significantly better 3-year OS (50% versus 41%, P = 0.005) [6]. Toxicity, while tolerable, was substantial; 41% and 32% of patients had grade 3 or 4 toxicity, respectively, including three (1%) treatment-related deaths. This treatment is now considered a standard care option for R0 patients with locally advanced disease [7–9]. Over half the patients in Intergroup-116 had less than a D1 dissection. Data from two large randomized trials comparing D1 and D2 nodal dissections, without adjuvant treatment, suggest that outcome following D1 or D2 dissections may resemble that achieved with adjuvant chemoradiation [35, 36]. Whether or not more adequate surgery would yield a similar improvement as in Intergroup-116, it would still apply for only a minority of patients. The primary inclusion criterion in this study was an R0 resection. According to the US National Cancer Database, this is achieved in only 30% of all gastric cancer patients [5]. Understanding the factor of patient selection, the results of earlier neoadjuvant trials and ours are encouraging, as is the 45% 3-year survival we noted. Further improvements in adjuvant and neoadjuvant therapy will probably require development of more effective chemotherapeutic regimens. During the last decade, several new agents with promising activity against gastric cancer were identified. These include paclitaxel, docetaxel and irinotecan [37, 38]. These agents are now undergoing phase II and III trials, as part of combination regimens. If improved outcome is seen in advanced disease, these agents will undergo extensive testing in the adjuvant setting. Pilot trials are already under way. Response to neoadjuvant therapy has been shown, in several models (e.g. breast, lung and rectal cancers), to predict
doi:10.1093/annonc/mdl133 | 1409
Downloaded from http://annonc.oxfordjournals.org/ at Bibliotheque de l'Universite Laval on July 12, 2015
quality of life assessment The number of quality of life questionnaires completed exhibited a sharp decrease over time, with a compliance rate of 40%–70%. There was an increasing trend in the total score until period 4 (6 months), followed by a slight decrease between periods 4 and 7 (6–15 months). Understanding that the number of patients at risk was small and noting the compliance drop in later time points, no significant change in quality of life during treatment was suggested.
original article
original article
acknowledgements The authors wish to thank Aminry Reyna for her assistance in collecting the data. The study was supported in part by Grant no. RO1 CA56225 from the National Cancer Institute, Bethesda, MD.
references 1. Parkin DM, Bray FI, Devesa SS. Cancer burden in the year 2000. The global picture. Eur J Cancer 2001; 37 (Suppl 8): S4–66. 2. Jemal A, Murrey T, Samuels A et al. Cancer statistics, 2003. CA Cancer J Clin 2003; 53: 5–26. 3. Blot WJ, Devesa SS, Kneller RW, Fraumeni JF. Rising incidence of adenocarcinoma of the esophagus and gastric cardia. J Am Med Assoc 1999; 265: 1287–1289. 4. Hermanek P, Maruyama K, Sobin LH. Stomach Cancer. In Hermanek P, Gospodarowicz MK, Henson D et al. (eds): Prognostic Factors in Cancer. Berlin: Springer 1995. 5. Hundahl S, Phillips J, Menck H. The National Cancer Data Base report on poor survival of U.S. gastric carcinoma patients treated with gastrectomy: fifth edition American Joint Committee on Cancer staging, proximal disease, and the ‘different disease’ hypothesis. Cancer 2000; 88: 921–932. 6. MacDonald J, Smalley S, Benedetti J et al. Chemoradiotherapy after surgery compared with surgery alone for adenocarcinoma of the stomach or gastroesophageal junction. N Engl J Med 2001; 345: 725–730. 7. Kelsen DP. Postoperative adjuvant chemoradiation therapy for patients with resected gastric cancer: Intergroup 116. J Clin Oncol 2000; 18: 32s–34s. 8. Smalley S, Gunderson L, Tepper J et al. Gastric surgical adjuvant radiotherapy consensus report: rationale and treatment implementation. Int J Radiation Oncology Biol Phys 2002; 52: 283–293. 9. Shah MA, Kelsen DP. Post-operative adjuvant chemoradiotherapy in high risk gastric cancer. Der Chirurg 2002; 73: 325–330. 10. Cunningham D, Allum WH, Stenning SP, Weeden S. Perioperative chemotherapy in operable gastric and lower oesophageal cancer: final results of a randomised, controlled trial (the MAGIC trial, ISRCTN 93793971). J Clin Oncol 2005; 23 (16S, Part I of II): 308S. 11. Gunderson LL SH. Adenocarcinoma of the stomach: areas of failure in a reoperation series (second or symptomatic look) clinicopathologic correlation and implications for adjuvant therapy. Int J Radiation Oncol Biol Phys 1992; 8: 1–11. 12. Dedrick R. Theoretical and experimental bases of intraperitoneal chemotherapy. Semin Oncol 1985; 12 (3 Suppl 4): 1–6.
1410 | Brenner et al.
13. Speyer JL, Sugarbaker PH, Collins JM et al. Portal levels and hepatic clearance of 5-fluorouracil after intraperitoneal administration in humans. Cancer Res 1981; 41: 1916–1922. 14. Kelsen DP SL, Cohen AM, Yao TJ et al. Related articles. A phase I trial of immediate postoperative intraperitoneal floxuridine and leucovorin plus systemic 5-fluorouracil and levamisole after resection of high risk colon cancer. Cancer 1994; 74: 2224–2233. 15. Alberts DS, Markman M, Armstrong D et al. Intraperitoneal therapy for stage III ovarian cancer: a therapy whose time has come! J Clin Oncol 2002; 20: 3944–3946. 16. Kelsen DP. Adjuvant and neoadjuvant therapy for gastric cancer. Semin Oncol 1996; 23: 379–389. 17. Vanhoefer U, Rougier P, Wilke H et al. Final results of a randomized phase III trial of sequential high-dose methotrexate, fluorouracil, and doxorubicin versus etoposide, leucovorin, and fluorouracil versus infusional fluorouracil and cisplatin in advanced gastric cancer: a trial of the European Organization for Research and Treatment of Cancer Gastrointestinal Tract Cancer Cooperative Group. J Clin Oncol 2000; 18: 2648–2657. 18. Ensminger WD, Rosowsky A, Raso V et al. A clinical-pharmacological evaluation of hepatic arterial infusions of 5-fluoro-29-deoxyuridine and 5-fluorouracil. Cancer Res 1978; 38 (11 part 1): 3784–3792. 19. Beahrs OH HD, Hutter RVP, Kennedy BJ (eds). Manual for Staging of Cancer. American Joint Committee on Cancer. Philadelphia: J.B. Lippincott 1992. 20. Fleming I, Cooper J, Henson D et al. American Joint Committee on Cancer Staging Manual. Philadelphia: Lippincott-Raven 1997. 21. Ruo L, Tickoo S, Klimstra DS et al. Long-term prognostic significance of extent of rectal cancer response to preoperative radiation and chemotherapy. Ann Surg 2002; 236: 75–81. 22. Rusch V, Klimstra D, Venkatraman E et al. Aberrant p53 expression predicts clinical resistance to cisplatin-based chemotherapy in locally advanced non-small cell lung cancer. Cancer Res 1995; 55: 5038–5042. 23. Kaplan EL, Meier P. Non-parametric estimation from incomplete observation. J Am Stat 1958; 53: 457–481. 24. Mantel N. Evaluation of survival data and two new rank order statistics arising in its consideration. Cancer Chemother Rep 1966; 50: 163–170. 25. Lin LI. A concordance correlation coefficient to evaluate reproducibility. Biometrics 1989; 45: 255–268. 26. Allum W, Cunningham D, Weeden S et al. Perioperative chemotherapy and operable gastric and lower esophageal cancer: a randomized, controlled trial (the MAGIC trial ISRCNT93793971). Am Soc Clin Oncol 2003; 22: 249. 27. Estape J, Grau JJ, Lcobendas F et al. Mitomycin C as an adjuvant treatment to resected gastric cancer. A 10-year follow-up. Ann Surg 1991; 213: 219– 221. 28. Laundry J, Tepper J, Wood W et al. Patterns of failure following curative resection of gastric carcinoma. Int J Radiation Oncol Biol Phys 1990; 19: 1357– 1362. 29. Lowy AM, Mansfield PF, Leach SD et al. Response to neoadjuvant chemotherapy best predicts survival after curative resection of gastric cancer. Ann Surg 1999; 229: 303–308. 30. Crookes P, Leichman CG, Leichman L et al. Systemic chemotherapy for gastric carcinoma followed by postoperative intraperitoneal therapy: a final report. Cancer 1997; 79: 1767–1775. 31. Siewert JR, Bottcher K, Roder JD et al. Prognostic relevance of systemic lymph node dissection in gastric carcinoma. Br J Surg 1993; 80: 1015–1018. 32. Kelsen D, Karpeh M, Schwartz G, et al. Neoadjuvant therapy of high-risk gastric cancer: a phase II trial of preoperative FAMTX and postoperative intraperitoneal fluorouracil-cisplatin plus intravenous fluorouracil. J Clin Oncol 1996; 14: 1818. 33. Alexander HR, Grem JL, Hamilton JM et al. Thymidylate synthase protein expression association with response to neoadjuvant chemotherapy and resection for locally advanced gastric and gastroesophageal adenocarcinoma. Cancer J Sci Am 1995; 1: 49. 34. Kang YK, Choi DW, Im YH et al. A phase III randomized comparison of neoadjuvant chemotherapy followed by surgery versus surgery for locally advanced stomach cancer. Proc ASCO 1996; 15: 215.
Volume 17 | No. 9 | September 2006
Downloaded from http://annonc.oxfordjournals.org/ at Bibliotheque de l'Universite Laval on July 12, 2015
improved patient outcome [21, 39, 40]. Mandard et al. [41] were the first to describe a method for pathologic response assessment in upper gastrointestinal malignancies which successfully predicted outcome. In their study, on esophageal cancer patients receiving preoperative chemoradiation, pathologic response strongly correlated with DFS. A comparable method was recently evaluated in gastric cancer [29]. Using a similar approach, our results suggest that pathologic response is associated with improved outcome in gastric cancer. If confirmed in larger trials, pathologic response would allow rapid evaluation of efficacy for future neoadjuvant studies. In summary, neoadjuvant cisplatin/5-FU and postoperative i.p. FUdR/LV can be safely delivered to patients undergoing radical gastrectomy and extensive lymph node dissection. The R0 resection rate is high, with no increase in operative morbidity and mortality. The failure pattern is consistent with possibly lower rate of i.p. and hepatic failures. New trials, using the same approach and employing possibly more effective chemotherapy regimens, are underway.
Annals of Oncology
Annals of Oncology
35. Bonenkamp JJ, Songun J, Hermans J et al. Randomised comparison of morbidity after D1 and D2 dissection for gastric cancer in 996 Dutch patients. Lancet 1995; 345: 745–748. 36. Cuschieri A, Weeden S, Fielding J et al. Patient survival after D1 and D2 resections for gastric cancer: long-term results of the MRC randomized surgical trial. Surgical Cooperative group. Br J Cancer 1999; 79: 1522–1530. 37. Shah MA. Recent developments in the treatment of gastric carcinoma. Curr Oncol Rep 2002; 4: 213–221. 38. Janunger KG, Hafstrom L, Nygren P, Glimelius B. A systemic overview of chemotherapy effects in gastric cancer. Acta Onc 2001; 40: 309–326.
original article 39. Bonadonna G, Valagussa P, Brambilla C et al. Primary chemotherapy in operable breast cancer: eight-year experience at the Milan Cancer Institute. J Clin Oncol 1998; 16: 93–100. 40. Pisters KM, Kris MG, Gralla RJ et al. Pathologic complete response in advanced non-small-cell lung cancer following preoperative chemotherapy: implications for the design of future non-small-cell lung cancer combined modality trials. J Clin Oncol 1993; 11: 1757–1762. 41. Mandard AM, Dalibard F, Mandard JC et al. Pathologic assessment of tumor regression after preoperative chemoradiotherapy of esophageal carcinoma: clinicopathologic correlations. Cancer 1994; 73: 2680–2686.
Downloaded from http://annonc.oxfordjournals.org/ at Bibliotheque de l'Universite Laval on July 12, 2015
Volume 17 | No. 9 | September 2006
doi:10.1093/annonc/mdl133 | 1411