- Email: [email protected]
The relationship of pathologic tumor regression grade (TRG) and outcomes after preoperative therapy in rectal cancer
Int. J. Radiation Oncology Biol. Phys., Vol. 62, No. 3, pp. 752–760, 2005 Copyright © 2005 Elsevier Inc. Printed in the USA. All rights reserved 0360-3016/05/$–see front matter
doi:10.1016/j.ijrobp.2004.11.017
CLINICAL INVESTIGATION
Rectum
THE RELATIONSHIP OF PATHOLOGIC TUMOR REGRESSION GRADE (TRG) AND OUTCOMES AFTER PREOPERATIVE THERAPY IN RECTAL CANCER FABIO MARIA VECCHIO, M.D.,* VINCENZO VALENTINI, M.D.,† BRUCE D. MINSKY, M.D.,‡ GILBERT D. A. PADULA, M.D.,‡ ENNAPADAM S. VENKATRAMAN, PH.D.,§ MARIO BALDUCCI, M.D.,† FRANCESCO MICCICHÈ, M.D.,† RICCARDO RICCI, M.D.,* ALESSIO GIUSEPPE MORGANTI, M.D.,† MARIA ANTONIETTA GAMBACORTA, M.D.,† FRANCESCA MAURIZI, M.D.,† AND CLAUDIO COCO, M.D.储 Departments of *Pathology and 储Surgery and †Division of Radiotherapy, Catholic University of the Sacred Heart, Rome, Italy; Departments of ‡Radiation Oncology and §Epidemiology and Biostatistics, Memorial Sloan-Kettering Cancer Center, New York, NY Purpose: To examine the relationship between tumor regression grade (TRG) and outcomes in patients with rectal cancer treated with preoperative therapy. Methods and Materials: Specimens from 144 patients with cT3,4 rectal cancer who had received preoperative radiation ⴞ chemotherapy and had a minimum follow-up of 3 years were retrospectively reviewed. TRG, which involves examining the residual neoplastic cells and scoring the degree of both cytological changes, including nuclear pyknosis or necrosis and/or eosinophilia, as well as stromal changes, including fibrosis (either dense or edematous) with or without inflammatory infiltrate and giant-cell granulomatosis around ghost cells and keratin, was quantified in five grades according to the Mandard score (Cancer 1994;73: 2680 –2686). The greater the response, the lower the TRG score. The median follow-up was 72 months (range, 40 –143 months). Results: Of the 144 patients, 19% were TRG1, 12% were TRG2, 21% were TRG3, 46% were TRG4, and 1% were TRG5. To simplify the analysis, TRG was combined into two groups: TRG1–2 and TRG3–5. By univariate analysis, none of the pretreatment factors examined, including age, circumference, length, distance from the anorectal ring, pretreatment T and N stage, and INDpre (defined as the pretreatment reference index size based on digital rectal examination), had an impact on 5-year outcomes, including local control, metastases-free survival, disease-free survival, and overall survival. Postoperative parameters, including pathologic T stage (pT), pathologic N stage (pN), and TRG, did significantly influence 5-year outcomes. These included local failure: pT0 –2: 5% vs. pT3– 4: 19%, p ⴝ 0.007; pN0: 7% vs. pN1–3: 26%, p ⴝ 0.002; TRG1–2: 2% vs. TRG3–5: 17%, p ⴝ 0.013; metastasis-free survival: pT0 –2: 86% vs. pT3– 4: 62%, p ⴝ 0.005; pNⴚ: 86% vs. pNⴙ: 42%, p < 0.001; TRG1–2: 91% vs. TRG3–5: 66%, p ⴝ 0.004; disease-free survival: pT0 –2: 83% vs. pT3– 4: 54%, p ⴝ 0.001; pN0: 80% vs. pN1–3: 39%, p < 0.001; TRG1–2: 91% vs. TRG3–5: 58%, p < 0.001; and overall survival: pT0 –2: 85% vs. pT3– 4: 65%, p ⴝ 0.007; pN0: 86% vs. pN1–3: 45%, p < 0.001; TRG1–2: 89% vs. TRG3–5: 68%, p ⴝ 0.004. By multivariate analysis combining all pre- and posttreatment parameters, only pN (p < 0.001) and TRG (p ⴝ 0.005) significantly predicted disease-free survival. Furthermore, TRG predicted the incidence of pathologic nodal involvement (p < 0.0001). Conclusions: By univariate analysis, TRG is a predictor for local failure, metastases-free survival, and overall survival. By multivariate analysis, it predicts improved disease-free survival. Given the ability of TRG to predict those patients with Nⴙ disease, it may be helpful, in combination with other clinicopathologic factors, in selecting patients for a more conservative procedure, such as local excision rather than radical surgery, after preoperative therapy. © 2005 Elsevier Inc. Neoadjuvant chemoradiation, Prognostic factors, Rectal neoplasms.
Reprint requests to: Vincenzo Valentini, M.D., Istituto di Radiologia, Università Cattolica del Sacro Cuore, Policlinico Universitario, “Agostino Gemelli,” Largo A. Gemelli, 8, 00168 Roma, Italy. Tel: (⫹39) 6-30154376; Fax: (⫹39) 6-35511338; E-mail: [email protected]
These data were presented during the poster discussion session at the 44th Annual Meeting of ASTRO in New Orleans, LA, on October 6 –10, 2002. Received July 13, 2004, and in revised form Oct 28, 2004. Accepted for publication Nov 8, 2004. 752
TRG in rectal cancer
INTRODUCTION In the last several years, a number of studies examining the use of preoperative chemoradiation (CMT) in patients with cT3– 4 rectal cancer have reported pathologic complete response rates (pCR) of 9 –29%, as well as an increased ability to perform sphincter-sparing surgery (1– 6). Many series report that patients who achieve a pCR after preoperative CMT, independent of their initial clinical T and N stage, have improved long-term outcomes, including local control, metastases-free survival, and overall survival (6 –12). In patients who achieve less than a pCR, there is heterogeneity in definitions and techniques of identifying and scoring the presence of residual tumor after preoperative CMT. For example, stage pT3 has been defined as responses ranging from gross disease remains in the perirectal fat to those with a few foci of microscopic residual disease outside the bowel wall. Most series report that the degree of response is predictive of outcomes (6, 10). For example, Wheeler et al. have suggested that patients with only microscopic foci in the mesorectum had a better prognosis compared with patients with T3 (transmural) disease (13). The aim of this study is to evaluate the relationship between the tumor regression grade (TRG) pathologic scoring system, which is based on the characteristics of the disposition of residual neoplastic cells in the specimen, and 5-year outcomes in patients with rectal cancer treated with preoperative radiation therapy with or without concomitant chemotherapy. METHODS AND MATERIALS A total of 216 patients with cT3– 4 and/or N⫹ rectal cancer were treated with preoperative radiation with or without concurrent chemotherapy between March 1990 and February 2000 in the Department of Radiation Oncology at the Catholic University, Rome, Italy. For this analysis, 144 patients were identified who had a minimum follow-up of 3 years and the availability of the pathologic specimen for the determination of TRG score. The remaining 72 patients were excluded because of inadequate follow-up (n ⫽ 5) or the lack of pathologic material for re-review (n ⫽ 67). The pathologic specimens were reviewed in serial section and evaluated to define the TRG by two independent pathologists (V.F.M., R.R.) who were not aware of the patient’s treatment protocol or outcome. In the rare cases of disagreement, the final assessment was achieved by the same pathologists using a multihead microscope. All patients were enrolled in previously reported Phase I/II trials using a variety of preoperative radiation or CMT regimens (4, 14 –18). The radiation therapy doses ranged from 37.8 Gy to 50.4 Gy at 1.8 Gy/fraction, and the chemotherapeutic agent(s) used included one or more of the following: 5-fluorouracil (5-FU), mitomycin C, cisplatin, oxaliplatin, and raltitrexed.
Eligibility criteria and diagnostic workup Eligibility criteria for all of the protocols were the same and included the following: the most inferior aspect of the primary tumor’s being between 0 and 10 cm from the anorectal ring as evaluated by double-contrast barium enema, histologically con-
● F. M. VECCHIO et al.
753
firmed adenocarcinoma, clinical stage T3–T4 and/or N1–N3, M0 according to the 1992 American Joint Committee on Cancer staging system (19), no prior radiation therapy to the pelvis, Eastern Cooperative Oncology Group performance status of 0 –2, granulocyte count greater than 3,000 per L, platelet count greater than 100,000 per L, hemoglobin concentration greater than 10 dL, serum creatinine value not greater than 1.5 mg/dL, creatinine clearance ⬎65 mL/min, no major intercurrent disease, adequate heart function, age ⬎18 years, and informed consent. Pretreatment workup included digital rectal examination, pelvic CT scan, transrectal ultrasound, liver ultrasonography, chest Xray, double-contrast barium enema, and proctoscopy with biopsy. Four to 5 weeks after the end of preoperative treatment, preoperative restaging was performed. The clinical response was determined during a weekly meeting of all specialists involved in the diagnostic investigation. Data from individual physicians’ examinations were compared, and the definitive combined staging and the tumor response were recorded (20). In the event of a disagreement, majority consensus was used. For the cN stage, any visible nodes on CT scan were considered positive, and patients with nodes identified in the region of the inferior mesenteric artery were considered N3.
Treatment schedules The patients were treated on one of six different preoperative sequential protocols performed between 1990 and 2000. 1. T3-FUMIR: 40 patients (28%) received bolus i.v. mitomycin C (10 mg/m2, Day 1) plus continuous infusion 5-FU (1,000 mg/m2, Days 1– 4), combined with 37.8 Gy (14). 2. T3-IORT: 23 patients (16%) received 37.8 Gy, and at surgery a 10 Gy intraoperative boost (IORT) was delivered to the presacral region with 6 MeV electrons (15). 3. T3-PLAFUR: 41 patients (25%) were treated with a 1– 4-h infusion of cisplatin (60 mg/m2, Days 1, 29) plus continuous infusion 5-FU (1,000 mg/m2, Days 1– 4/5 and 29 –32/33) and concurrent 45 Gy followed by a boost to 50.4 Gy (4). 4. T3-TOMRT: 7 patients (5%) received raltitrexed (3 mg/m2 Days 1, 19, 38) plus concurrent 45 Gy followed by a boost to 50.4 Gy. At surgery, a 10 Gy IORT boost was delivered to the presacral region with 6 MeV electrons (16). 5. T3-TOMOXRT: 9 patients (6%) received raltitrexed (3 mg/ m2) and dose escalation with oxaliplatin (60 – 85–110 –130 mg/m2 Days 1, 19, 38) concurrent with 45 Gy followed by a boost to 50.4 Gy (17). 6. T4-FUMIR: 24 patients (17%) received bolus i.v. mitomycin C (10 mg/m2, Day 1) plus continuous infusion 5-FU (1,000 mg/m2 Weeks 1, 5) concurrent with 45– 48 Gy. At surgery, 4 patients received a 10 Gy IORT boost to the presacral region with 6 MeV electrons (18). Overall, 23 patients (16%) were treated with preoperative radiotherapy alone, 121 (84%) with preoperative CMT, and 29 (20%) with postoperative adjuvant chemotherapy (Table 1).
Radiation therapy techniques The clinical target volume 2 (CTV2) included the tumor, the mesorectum, and the internal iliac nodes, with or without external iliac lymph nodes. A 4-field or 3-field technique was used. The lateral border of anteroposterior-posteroanterior radiation fields was 1.5–2 cm outside the true bony pelvis; the inferior border was 1 cm above the anal verge in the tumors of the middle rectum and
754
I. J. Radiation Oncology
● Biology ● Physics
Table 1. Patient characteristics No. Age (years) Median (range) 64 (25–81) Gender Male Female cTNM stage T2N0 T2N1 T2N2 T3N0 T3N1 T3N2 T3N3 T4N0 T4N1 T4N2 T4N3 Treatment Preoperative radiotherapy alone Preoperative chemoradiation Adjuvant chemotherapy Histologic classification Adenocarcinoma G1–G2 Adenocarcinoma G3 Mucinous adenocarcinoma Adenocarcinoma NOS
%
Volume 62, Number 3, 2005
scopic residual disease. IORT was delivered using 6 MeV electrons, and the dose was 10 Gy.
Postoperative chemotherapy
96 48
66.6 33.3
1 3 1 23 52 38 3 1 6 13 3
0.7 2.1 0.7 15.9 36.1 26.3 2.1 0.7 4.1 9 2.1
23 121 29
15.9 84.1 20.1
65 13 2 63
45 9 1.4 44.4
just below the anal verge in the tumors of the lower rectum; the superior border was at least 2 cm above the tumor and not inferior to the sacral promontory; corner blocks or multileaf collimators were used to exclude extrapelvic normal tissues. The lateral field posterior border was a minimum of 1.5 cm behind the anterior bony sacral margin and the anterior border at the most posterior aspect of the symphysis pubis. In patients with T4 disease, this border was anterior to the symphysis pubis to include external iliac nodes. The clinical target volume 1 (CTV1) included the mesorectum corresponding to the tumor mass with 2-cm radial margins. Radiation was delivered with a 10 MV linear accelerator. Fractionation was at 180 cGy/day, 5 fractions per week. All fields were treated daily. The dose was prescribed to the axis intersection according to ICRU 50/62 (21, 22).
Surgery Surgery was performed 4 to 8 weeks after the completion of radiation. The choice of the surgical procedure (abdominoperineal resection [APR], low anterior resection [LAR], or local excision) was at the surgeon’s discretion. Temporary colostomy after LAR was performed at the surgeon’s discretion. For patients who underwent an LAR or APR, the resection of whole mesorectum (total mesorectal excision) was performed and inked distal and radial margins were evaluated. For those undergoing an LAR, a distal rectal margin of at least 2 cm was obtained.
Intraoperative radiation therapy A total of 34 patients with cT3 (n ⫽ 30) or cT4 (n ⫽ 4) disease were referred by a team of surgeons involved in a pilot study of IORT (20). The IORT target was chosen according to the size of the resected tumor and of the area thought to be at risk of micro-
All patients with stage cT4 (pretherapy), as well as those on protocols T3-TOMRT and T3-TOMOXRT with pN⫹ (n ⫽ 29), received postoperative chemotherapy as part of their primary treatment. The regimen was bolus 5-FU (400 mg/m2/day) and leucovorin (100 mg/m2 Days 1–5) monthly for 6 months (23).
Follow-up Patients were seen in routine follow-up every 3 months for the first year, every 6 months for the next 2 years, and then yearly. At each follow-up visit, a digital examination and proctoscopy were performed. Liver ultrasound, chest X-ray, pelvic CT, and carcinoembryonic antigen were performed at 3–12-month intervals (20). The median follow-up was 72 months (range: 40 –143 months). No patient was lost to follow-up.
Evaluation of response The tumor response was assessed according to the World Health Organization score (24). Clinical response on digital rectal examination was also evaluated using a reference index (IND) defined as the product of the quarters of circumference of rectal wall involved (categorized 1 to 4) multiplied by the craniocaudal length of the tumor in mm (4). This index was calculated at the time of initial diagnosis (INDpre) and 4 to 6 weeks after radiation (INDpost). INDpre was calculated in all patients, and INDpost was available in 97 patients. Downstaging was defined as any reduction of stage comparing clinical T (cT) and N (cN) stage to pathologic T (pT) and N (pN) stage.
Tumor regression grade Tumor regression grade (TRG) as reported by Mandard et al. in patients treated for esophageal cancer was used to assess the pathologic tumor response after preoperative therapy (25). Tumor regression after preoperative therapy was assessed by examining the residual neoplastic cells and scoring the degree of both cytological changes, including nuclear pyknosis or necrosis and/or eosinophilia, and stromal changes, including fibrosis (either dense or edematous) with or without inflammatory infiltrate and giantcell granulomatosis around ghost cells and keratin. On the basis of the combination of these changes, tumor regression was classified according to the five following grades (Fig. 1): TRG1: Complete response with absence of residual cancer and fibrosis extending through the wall. TRG2: Presence of residual cancer cells scattered through the fibrosis. TRG3: Increase in the number of residual cancer cells, with fibrosis predominant. TRG4: Residual cancer outgrowing fibrosis. TRG5: Absence of regressive changes.
Statistical analysis The distribution of ordered categories was analyzed by Gamma, Kendall’s, and Stuart’s tau tests (26). When one categorical variable was dichotomous and the other ordered, the test for linear trend of proportions was used (27). Differences of p ⱕ 0.05 were considered statistically significant. Local survival and disease-free and overall survival were estimated according to the Kaplan-Meier
TRG in rectal cancer
actuarial method. Statistical significance of each observed variable as a predictor of outcomes was analyzed by the Cox proportional hazard model (28). The clinical end points evaluated were local control, metastases-free survival, disease-free survival, and overall survival. Differences of p ⱕ 0.05 were considered statistically significant.
RESULTS Patient characteristics (pretreatment) A total of 144 patients (96 male, 48 female) with a minimum follow-up of 3 years had TRG analysis of pathologic specimen. The median age was 64 years (range, 25– 81 years). The tumor stage at initial (pretreatment) diagnosis and tumor grade are seen in Table 1. The median length of the tumor was 50 mm (SD, 17.6, range, 10 –150 mm). The number of quarters of rectal wall invaded by the tumor were measured on CT scan images; the results are as follows: 1 wall, 8 (6%); 2 walls, 69 (48%); 3 walls, 17 (12%); and 4 walls (circumferential), 50 (35%). IND was calculated as the product of number of quarters infiltrated by the tumor and the length of the tumor. The median INDpre was 120 mm (SD, 92.9, range, 25– 600 mm). The location of the tumor was defined by measuring the distance of the most inferior aspect of the tumor from the anorectal ring as seen on barium enema. The median distance was 40 mm (SD, 26.6; range, 12.5–120 mm). In 56 patients (39%) the tumor was located between 0 and 30 mm, in 43 patients (30%) it was located between 30 and 50 mm, and in 45 patients (31%) it was more than 50 mm. Response Four to 6 weeks after the completion of preoperative therapy, all patients were reevaluated with interval history and physical examination, CT scan, and barium enema. Tumor response comparing INDpre with INDpost was assessed using World Health Organization criteria. The overall clinical response rate was 51%: 4% clinical complete response and 47% partial response; the majority (49%) had a minor response that was classified as no change. No patients developed disease progression. INDpost was available in 97 patients; it ranged from 10 to 280 with a median value of 40 (SD, 54.7). Surgical procedures All patients underwent surgery after preoperative therapy; the procedures included local excision: 2 (1%), LAR: 82 (57%), LAR plus IORT: 11 (8%), APR: 29 (20%), APR plus IORT: 18 (13%), and Hartmann procedure: 2 (1%). Sphincter-sparing surgery was performed in 66% of patients. In the 56 patients in whom the most inferior aspect of the tumor was 0 –30 mm from the anorectal ring, 17 (30%) underwent sphincter-sparing surgery. Downstaging Posttreatment pathologic stages included pT0N0: 25 (17%), pT0N1: 2 (1%), pT1N0: 8 (6%), pT1N1: 1 (1%),
● F. M. VECCHIO et al.
755
pT2N0: 21 (15%), pT2N1: 3 (2%), pT2N2: 1 (1%), pT3N0: 44 (31%), pT3N1: 15 (10%), pT3N2: 12 (8%), pT3N3: 3 (2%), pT4N0: 3 (2%), and pT4N1: 6 (4%). Overall, the percentage of downstaging was 53% (77/144) for T stage 77% and 78% (93/119) for N stage. The tumor regression grade was TRG1: 27 (19%), TRG2: 18 (13%), TRG3: 31 (22%), TRG4: 67 (47%), and TRG5: 1 (1%). To simplify the analysis, TRG was combined into two groups, TRG1–2 and TRG3–5. The relationship between TRG and pN was analyzed. Overall, 41 patients (28%) with TRG1–2 were pN0, and 4 (3%) with TRG1–2 were pN1–3. In contrast, 60 patients (42%) with TRG3–5 were pN0, and 39 (27%) with TRG3–5 were pN1–3, p ⬍ 0.0001 (Table 2). Therefore, 91% (41/45) of patients with TRG1–2 were pN0, whereas 63% (60/95) of patients with TRG3–5 were pN0. Patterns of failure and survival A total of 43 patients died: 34 deaths were due to tumor, and 9 were due to intercurrent disease. Of the remaining 101 patients, 53 are without evidence of disease. Patterns of failure in the remaining 48 patients include the following: local failure: 9 (7%), distant metastasis: 29 (20%), and both local failure and distant metastasie: 10 (7%). The sites of distant metastases were as follows: liver: 22 (15%), lung: 10 (7%), brain: 2 (1%), bone: 2 (1%), and other nonspecified sites in 3 (2%). The 5-year local failure, metastasis-free survival, disease-free survival, and overall survival were 87%, 73%, 67%, and 74%, respectively. Outcome according to pretreatment prognostic factors Univariate analysis of the impact of age, tumor circumference, length and distance from the anorectal ring, cT and cN stage, and INDpre is seen in Table 3. Younger patients had a significantly higher incidence of local failure (⬍50 years: 33%, 51–70 years: 9%, and ⬎70 years: 0%, p ⫽ 0.001) and lower 5-year disease-free survival (⬍50 years: 45%, 51–70 years: 71%, and ⬎70 years: 88%, p ⫽ 0.012). Tumor circumference and length were not prognostic factors. Overall 5-year survival was significantly lower for more distal tumors (⬍30 mm: 68% vs. 31–50 mm: 89%, p ⫽ 0.029). However, there was no correlation with cT stage. Clinical nodal status significantly correlated with disease-free survival, metastasis-free survival, and overall survival. The 5-year disease-free survival by nodal status was cN0: 78%, cN1: 78%, cN2: 53%, and cN3: 50%, p ⫽ 0.049. The 5-year metastasis-free survival was cN0: 78%,
Table 2. Correlation between TRG and pN (p ⬍ 0.0001) TRG
pN0
pN1–3
TRG1–2 TRG3–5
41 (28%) 60 (42%)
4 (3%) 39 (27%)
Abbreviation: TRG ⫽ tumor regression grade.
756
I. J. Radiation Oncology
● Biology ● Physics
Volume 62, Number 3, 2005
Table 3. Univariate analysis of pretreatment parameters: Outcome at 5 years
Age Circumference Length Distance from the anorectal ring cT cN INDpre
Local failure
Metastasis-free survival
Disease-free survival
Overall survival
p ⫽ 0.001 NS NS NS NS NS NS
p ⬍ 0.05 NS NS NS NS p ⫽ 0.014 NS
p ⫽ 0.012 NS NS NS NS p ⫽ 0.049 NS
NS NS NS p ⫽ 0.029 NS p ⬍ 0.029 NS
Abbreviation: NS ⫽ nonsignificant.
cN1: 87%, cN2: 56%, and cN3: 67%, p ⫽ 0.014. The 5-year overall survival was cN0: 87%, cN1: 78%, cN2: 66%, and cN3: 50%, p ⫽ 0.029. The following 5-year outcomes decreased with an increasing number of positive nodes (cN0 –1 vs. cN2–3): local failure (9% vs. 17%, p ⫽ 0.316), metastasis-free survival (85% vs. 57%, p ⫽ 0.002), diseasefree survival (78% vs. 53%, p ⫽ 0.005), and overall survival (81% vs. 65%, p ⫽ 0.003). There was no correlation between INDpre and the 5-year local failure, metastasis-free survival, disease-free survival, and overall survival. By multivariate analysis, the following variables were independent factors for outcomes. Local failure: age (p ⫽ 0.006) and distance of tumor from the anorectal ring (p ⫽ 0.033); metastasis-free survival: distance of tumor from anorectal ring (p ⫽ 0.003), length of the tumor (p ⫽ 0.015), lymph node status (p ⫽ 0.045), and INDpre (p ⫽ 0.019); disease-free survival: age (p ⫽ 0.019), distance of tumor from the anorectal ring (p ⫽ 0.003), lymph node status (p ⫽ 0.040), and INDpre (p ⫽ 0.029); and overall survival: distance of tumor from the anorectal ring (p ⫽ 0.004) and lymph node status (p ⫽ 0.016). Outcomes according to posttreatment prognostic factors As seen in Table 4, by univariate analysis, except for the distance from the anorectal ring, all posttreatment factors (pT, pN, INDpost, TRG) significantly predicted outcomes. Compared with patients with INDpost ⬍30, those with INDpost ⬎30 had a significant decrease in 5-year diseasefree survival (64% vs. 76%, p ⫽ 0.025) and overall survival (67% vs. 86%, p ⫽ 0.043). There was a significant decrease in favorable outcomes with increasing pT stage. This included 5-year local failure (pT0 –1: 4% vs. pT3– 4: 19%, p ⫽ 0.007), metastasis-free survival (pT0 –1: 86%
vs. pT3– 4: 62%, p ⫽ 0.005), disease-free survival (pT0 –1: 83% vs. pT3– 4: 54%, p ⫽ 0.001), and overall survival (pT0 –1: 85% vs. pT3– 4: 65%, p ⫽ 0.007). Likewise, there was a significant decrease in outcomes with increasing pN stage. This included 5-year local failure (pN0: 7% vs. pN1–3: 26%, p ⫽ 0.002), metastasis-free survival (pN0: 86% vs. pN1–3: 42%, p ⬍ 0.001), 5-year disease-free survival (pN0: 80% vs. pN1–3: 40%, p ⬍ 0.001), and overall survival (pN0: 86% vs. pN1–3: 45%, p ⬍ 0.001). With the exception of local control, as TRG increased, the corresponding 5-year outcomes decreased. The cumulative rate of distant metastases by TRG was TRG1: 93%, TRG2: 90%, TRG3: 86%, and TRG4: 56% (p ⫽ 0.001). Similar results were seen for 5-year disease-free survival; TRG1: 96%, TRG2: 90%, TRG3: 74%, and TRG4: 52% (p ⫽ 0.002). The overall rate of death was TRG1: 11%, TRG2: 5%, TRG3:19%, and TRG4: 42% (p ⫽ 0.002); 5-year survival was TRG1: 91%, TRG2: 85%, TRG3: 79%, and TRG4: 63% (p ⫽ 0.016). When TRG was combined into two groups (TRG1–2 vs. TRG3–5), the differences in outcomes remained significant (Fig. 1). This included 5-year local failure: TRG1–2: 2% vs. TRG3–5: 17%, p ⫽ 0.013 (Fig. 2); metastasis-free survival: TRG1–2: 91% vs. TRG3–5: 66%, p ⫽ 0.004 (Fig. 3); disease-free survival: TRG1–2: 91% vs. TRG3–5: 58%, p ⬍ 0.001 (Fig. 4); and overall survival: TRG1–2: 89% vs. TRG3–5: 68%, p ⫽ 0.004 (Fig. 5). By multivariate analysis, a number of factors significantly predicted 5-year outcomes. These included local failure: pN (p ⫽ 0.020) and TRG (p ⫽ 0.035); metastasis-free survival: pT (p ⫽ 0.010), pN (p ⬍ 0.001), and TRG (p ⫽ 0.005); disease-
Table 4. Univariate analysis of posttreatment parameters: Outcomes at 5 years
pT pN TRG1–2 vs. TRG3–5 INDpost Abbreviation: NS ⫽ nonsignifant.
Local failure
Metastasis-free survival
p ⫽ 0.007 p ⫽ 0.002 p ⫽ 0.013 NS
p ⫽ 0.005 p ⬍ 0.001 p ⫽ 0.004 NS
Disease-free survival p p p p
⫽ ⬍ ⬍ ⫽
0.001 0.001 0.001 0.025
Overall survival p p p p
⫽ ⬍ ⫽ ⫽
0.007 0.001 0.004 0.043
TRG in rectal cancer
Fig. 1. Tumor regression grade (TRG).
free survival: pN (p ⬍ 0.001) and TRG (p ⫽ 0.003); and overall survival: pN (p ⬍ 0.001) and TRG (p ⫽ 0.042). Combined multivariate analysis of both pre- and posttreatment factors Multivariate analysis of all pre- and posttreatment factors revealed that local failure, pN (p ⫽ 0.024), and age (p ⫽ 0.025) were independent prognostic factors. These included metastasis-free survival: pT (p ⫽ 0.037), pN (p ⬍ 0.001), and TRG (p ⫽ 0.005); disease-free survival: pN (p ⬍ 0.001) and TRG (p ⫽ 0.011); overall survival: pN (p ⬍ 0.001).
● F. M. VECCHIO et al.
757
Morgan et al. tested the interobserver reliability using TRG to assess pathologic tumor response after preoperative CMT in 21 patients with T3–T4 rectal cancer. A TRG score of 1–3 was seen in 62%, and TRG4 –5 was seen in 38%. The pCR rate was 14%. In contrast to our results, they observed that pT and pN stages were not directly related to TRG. However, the interobserver reliability in scoring TRG was high. The authors did not analyze the impact of TRG grade on outcomes (29). Other grading systems to quantify tumor regression after preoperative treatment have been developed. Wheeler et al. retrospectively reviewed surgical specimens of 42 patients with T3/T4 rectal cancer treated with preoperative CMT (13). The specimens were analyzed by one pathologist, and the Rectal Cancer Regression Grade (RCRG) was measured. The authors defined three grades of regression: RCRG 1: either pCR or only microscopic foci of adenocarcinoma, RCRG 2: marked fibrosis with macroscopic tumor still present, and RCRG 3: a poor response with little or absence of fibrosis in the presence of abundant macroscopic tumor. Overall, 46% were RCRG 1, 40% were RCRG 2, and 14% were RCRG 3. The authors speculated that RCRG better defines regression of disease after CMT compared with the pT or pN stage. The subset of patients with pT3 microscopic foci had a more favorable prognosis compared with those with pT3 macroscopic disease. At Memorial Sloan-Kettering Cancer Center, selected clinicopathologic factors, including the extent of pathologic response to preoperative radiation or CMT, were retrospec-
DISCUSSION Retrospective data suggest that pathologic downstaging after preoperative radiotherapy with or without concomitant chemotherapy is associated with improved outcomes (6 –10, 12). Patients with pT0 –2 stage disease after preoperative therapy have local failure rates of 0 – 6% and 5-year survivals of 90 –100%. However, there is heterogeneity in the definitions as well as the techniques of identifying and scoring the presence of residual tumor after preoperative therapy. Therefore, we have analyzed response using the TRG score, which offers a more uniform method of assessment. The TRG score was developed by Mandard et al. to assess the response of preoperative therapy in patients with esophageal cancer. They examined 93 resected specimens of esophageal cancer from patients treated with preoperative CMT. TRG score predicted 3-year survival independent of tumor stage; TRG1: 61%, TRG2: 60%, TRG3: 27%, TRG4: 0%, and TRG5: 0%, p ⬍ 0.0001. By univariate analysis, TRG, pN stage, tumor size, and the amount of wall involvement by tumor were significantly correlated with diseasefree survival (p ⬍ 0.05). By multivariate analysis, only TRG (TRG1–3 vs. TRG4 –5) remained a significant predictor of disease-free survival (p ⬍ 0.001). The authors recommended that TRG should be considered when evaluating therapeutic results after preoperative CMT in esophageal cancer (25).
Fig. 2. Local control according to TRG1–2 vs. TRG3–5 (p ⫽ 0.013). TRG ⫽ tumor regression grade.
758
I. J. Radiation Oncology
● Biology ● Physics
Fig. 3. Metastases-free survival according to TRG1–2 vs. TRG3–5 (p ⫽ 0.004). TRG ⫽ Tumor regression grade.
tively analyzed to examine the impact on long-term recurrence-free survival in 69 patients with T3– 4 and/or pN⫹ primary rectal cancer. A 5-point grading scale was used to categorize pathologic response to preoperative therapy: 0: no response, 1: response ⱕ1/3, 2: response 1/3–2/3, 3: response ⱖ2/3 but ⬍ complete, and 4: complete response. With a median follow-up of 69 months, 5-year relapse-free survival was 79% and was significantly worse for patients with aggressive pathologic features and pN⫹. Risk ratios for relapse-free survival were 3.68 for the presence of aggressive pathologic features and 4.64 for node-positive rectal cancers. In patients with greater than 95% response to preoperative therapy, only 1 patient was dead from disease; another died of an unrelated cause, and the remaining patients were free of disease with a minimum follow-up of 47 months. However, the extent of response did not predict relapse-free survival (12). Rodel et al. examined the apoptotic index (AI), Ki-67, p-53, and p-2 immunohistochemically on pretreatment biopsies in 44 patients with T3/4 rectal cancer after preoperative CMT (30). The treatment response was assessed with a 5-grade system using histopathologic methods on the resected surgical specimen. The AI was obtained by dividing the number of apoptotic tumor cells by the total number of tumor cells multiplied by 100; the pathologic response of the tumor was defined according to a rectal cancer regression grading defined by Dworak et al. (31). This includes Grade 0: no regression, Grade 1: minimal regression, Grade 2: moderate regression, Grade 3: good regression,
Volume 62, Number 3, 2005
Grade 4: total regression. Tumors with Grades 3 or 4 response to CMT had significantly higher pretreatment levels of apoptosis than tumors with lower regression (Grades 0 –2). The pretreatment AI predicted the risk of 5-year local failure: 100% for AI greater than median vs. 74% for AI less than median (p ⫽ 0.08). Furthermore, AI predicted tumor regression. Tumors with Grades 3/4 response to CMT had significantly higher pretreatment levels of apoptosis (mean AI: 2.06%) than tumors with Grades 0 –2 response (AI: 1.44%, p ⫽ 0.003). A significant correlation was also found between pathologic tumor response and relapse-free survival (Grade 3/4: 72% vs. Grade 0/1/2: 28%, p ⫽ 0.004). When the analysis was limited to local control, the only pretreatment factor significantly correlated with this end point was AI. Moreover, AI significantly correlated with Ki-67 but not with p-53 and bcl2 status. The authors concluded that AI and tumor regression can be helpful in selecting preoperative therapy (30). In our experience, by univariate analysis, pretreatment parameters (age, circumference, length and distance from the anorectal ring, pretreatment T and N stage, and INDpre) were associated with some but not all 5-year outcomes. Posttreatment parameters (pT, pN, and TRG) affected all 5-year outcomes. By multivariate analysis, pooling all preand posttreatment parameters, only pN and TRG significantly predicted disease-free survival. Furthermore, TRG significantly predicted the presence of pN⫹ disease. We have previously reported the impact of pathologic staging and of tumor index after treatment (INDpost) in 165
Fig. 4. Disease-free survival according to TRG1–2 vs. TRG3–5 (p ⬍ 0.001). TRG ⫽ tumor regression grade.
TRG in rectal cancer
● F. M. VECCHIO et al.
759
analysis, INDpost ⬍30 (i.e., ⱕ1/4 circumference ⱕ30 mm length) was predictive of 5-year local control and survival. Our data support the role of the TRG score in predicting 5-year outcomes. In contrast to INDpost, TRG offers a more reliable interpretation of the tumor regression in the pathologic specimen, accounting for the presence of residual isolated cells in a well-defined group at a lower risk. Not only does it decrease the chance of misinterpretation of the pathologic T stage as described by Morgan et al. (29), but the interobserver analysis is also more reliable. It should be emphasized that in our prior series examining INDpost, patients with residual isolated cancer cells in the specimen were scored as pT1, pT2, or pT3 based on the location of residual cells in the rectal wall or in the mesorectum (10). As with other clinicopathologic scoring systems, our preliminary, retrospective data need to be validated in a larger, prospective trial. It must be emphasized that it is retrospective with unknown selection bias related to the exclusion of patients with lack of pathologic material for re-review, and variables such as the impact of chemotherapy, radiation doses, the interval between radiation and surgery, and the pretreatment stage may account for the differences in results reported. Fig. 5. Overall survival according to TRG1–2 vs. TRG3–5 (p ⫽ 0.004). TRG ⫽ tumor regression grade.
CONCLUSION
consecutive patients with locally advanced rectal cancer, treated with preoperative 5-FU– based CMT (4). Increasing pT stage was a significant predictor for 5-year survival (pT0: 91%, pT2: 81%, pT3: 66%, and pT4: 47% [p ⫽ 0.014]), as well as local failure (pT0: 0%, pT1: 17%, pT2: 12%, pT3: 21%, and pT4: 54% [p ⫽ 0.0012]). The pN status significantly predicted freedom from distant metastases and overall survival. By univariate and multivariate
Tumor regression grade seems to be a prognostic factor for disease-free survival in patients receiving preoperative therapy for rectal cancer. In addition, it is a prognostic factor for local failure, metastasis-free survival, and overall survival. Given the ability of TRG to predict those patients with N⫹ disease, in combination with other clinicopathologic factors, it may be helpful after preoperative therapy in selecting patients for a more conservative procedure, such as local excision rather than radical surgery.
REFERENCES 1. Rich TA, Skibber JM, Ajani JA, et al. Preoperative infusional chemoradiation therapy for stage T3 rectal cancer. Int J Radiat Oncol Biol Phys 1995;32:1025–1029. 2. Chari RS, Tyler DS, Anscher MS, et al. Preoperative radiation and chemotherapy in the treatment of adenocarcinoma of the rectum. Ann Surg 1995;221:778 –786. 3. Grann A, Minsky BD, Cohen AM, et al. Preliminary results of preoperative 5-fluorouracil, low-dose Leucovorin and concurrent radiation therapy for clinically resectable T3 rectal cancer. Dis Colon Rectum 1997;40:515–522. 4. Valentini V, Coco C, Cellini N, et al. Ten years of preoperative chemoradiation for extraperitoneal T3 rectal cancer: Acute toxicity, tumor response, and sphincter preservation in three consecutive studies. Int J Radiat Oncol Biol Phys 2001; 51:371–383. 5. Bosset JF, Magnin V, Maingon P, et al. Preoperative radiochemotherapy in rectal cancer: Long-term results of a Phase II trial. Int J Radiat Oncol Biol Phys 2000;46:323–327. 6. Mohiuddin M, Hayne M, Regine WF, et al. Prognostic significance of postchemoradiation stage following preoperative chemotherapy and radiation for advanced/recurrent
7.
8.
9.
10.
11.
rectal cancers. Int J Radiat Oncol Biol Phys 2000;48: 1075–1080. Janjan NA, Abbruzzese J, Pazdur R, et al. Prognostic implications of response to preoperative infusional chemoradiation in locally advanced rectal cancer. Radiother Oncol 1999;51: 153–160. Berger C, de Muret A, Garaud P, et al. Preoperative radiotherapy (RT) for rectal cancer: Predictive factors of tumor downstaging and residual tumor cell density (RTCD): Prognostic implications. Int J Radiat Oncol Biol Phys 1997;37: 619 – 627. Kaminsky-Forrett MC, Conroy T, Luporsi E, et al. Prognostic implications of downstaging following preoperative radiation therapy for operable T3-T4 rectal cancer. Int J Radiat Oncol Biol Phys 1998;42:935–941. Valentini V, Coco C, Picciocchi A, et al. Does downstaging predict improved outcome after preoperative chemoradiation for extraperitoneal locally advanced rectal cancer? A longterm analysis of 165 patients. Int J Radiat Oncol Biol Phys 2002;53:664 – 674. Gerard JP, Chapet O, Morignat E, et al. Preoperative radio-
760
12. 13. 14.
15.
16.
17.
18. 19. 20.
I. J. Radiation Oncology
● Biology ● Physics
therapy of rectal cancer. The Lyons experience, 1985–1996. Prognostic study apropos of 312 patients. Ann Chir 1999;53:1003–1010. 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. Wheeler JM, Warren BF, Mortensen NJMcC, et al. Quantification of histologic regression of rectal cancer after irradiation. Dis Colon Rectum 2002;45:1051–1056. Valentini V, Coco C, Cellini N, et al. Preoperative chemoradiation for extraperitoneal T3 rectal cancer: Acute toxicity, tumor response, sphincter preservation. Int J Radiat Oncol Biol Phys 1998;40:1067–1075. Pacelli F, Di Giorgio A, Papa V, et al. Preoperative radiotherapy combined with intraoperative radiotherapy improves results of total mesorectal excision in patients with T3 rectal cancer. Dis Colon Rectum 2004;47:170 –179. Valentini V, Doglietto GB, Morganti AG, et al. Preoperative chemoradiation with raltitrexed (‘Tomudex’) for T2/N ⫹ and T3/N ⫹ rectal cancers: A phase I study. Eur J Cancer 2001; 37:2050 –2055. Valentini V, Morganti AG, Luzi S, et al. Chemoradiation with raltitrexed (‘Tomudex’)and oxaliplatin in pre-operative treatment of stage II/III resectable rectal cancer: A dose finding study. Proc ASCO 2001;20:131a. Sofo L, Ratto C, Doglietto GB, et al. Intraoperative radiation therapy in integrated treatment of rectal cancers. Results of phase II study. Dis Colon Rectum 1996;39:1396 –1403. UICC International Union Against Cancer. TNM classification of malignant tumours. 4th ed., 2nd rev. Springer Verlag; 1992. Marano P, Barbaro B, De Franco A, et al. Diagnostic and therapeutic integration in rectal cancer. Preliminary notes. Radiol Med (Torino) 1992;84:261–273.
Volume 62, Number 3, 2005
21. International Commission on Radiation Units and Measurements. ICRU Report 50: Prescribing, recording and reporting photon beam therapy. 1993. 22. International Commission on Radiation Units and Measurements. ICRU Report 62: Prescribing, recording and reporting photon beam therapy (supplement to ICRU Report 50). ICRU News, December 1999. 23. Machover D, Schwarzenberg L, Goldschmidt E, et al. Treatment of advanced colorectal and gastric adenocarcinomas with 5-FU combined with high-dose folinic acid: A pilot study. Cancer Treat Rep 1982;66:1803–1807. 24. Miller AB. Reporting result of cancer treatment. Cancer 1981; 47:207–214. 25. Mandard AM, Dalibard F, Mandard JC, et al. Pathologic assessment of tumor regression grade after preoperative chemoradiotherapy of esophageal carcinoma. Cancer 1994;73: 2680 –2686. 26. Brown MB, Benedetti JK. Sampling behavior of tests for correlation in two-way contingency tables. J Am Stat Assoc 1977;72:309 –315. 27. Cochran WG. Some methods for strengthening the common chi2 tests. Biometrics 1954;10:417– 441. 28. Cox DR. Regression model and life tables. J R Stat Soc Series B 1972;34:187–220. 29. Morgan MJ, Koorey DJ, Painter D, et al. Histological tumour response to pre-operative combined modality therapy in locally advanced rectal cancer. Colorectal Dis 2002;4:177–183. 30. Rodel C, Grabenbauer GG, Papadopoulos T, et al. Apoptosis as cellular predictor for histopathologic response to neoadjuvant radiochemotherapy in patients with rectal cancer. Int J Radiat Oncol Biol Phys 2002;52:294 –303. 31. Dworak O, Keilholz L, Hoffmann A. Pathological features of rectal cancer after preoperative radiochemotherapy. Int J Colorectal Dis 1997;12:19 –23.
doi:10.1016/j.ijrobp.2004.11.017
CLINICAL INVESTIGATION
Rectum
THE RELATIONSHIP OF PATHOLOGIC TUMOR REGRESSION GRADE (TRG) AND OUTCOMES AFTER PREOPERATIVE THERAPY IN RECTAL CANCER FABIO MARIA VECCHIO, M.D.,* VINCENZO VALENTINI, M.D.,† BRUCE D. MINSKY, M.D.,‡ GILBERT D. A. PADULA, M.D.,‡ ENNAPADAM S. VENKATRAMAN, PH.D.,§ MARIO BALDUCCI, M.D.,† FRANCESCO MICCICHÈ, M.D.,† RICCARDO RICCI, M.D.,* ALESSIO GIUSEPPE MORGANTI, M.D.,† MARIA ANTONIETTA GAMBACORTA, M.D.,† FRANCESCA MAURIZI, M.D.,† AND CLAUDIO COCO, M.D.储 Departments of *Pathology and 储Surgery and †Division of Radiotherapy, Catholic University of the Sacred Heart, Rome, Italy; Departments of ‡Radiation Oncology and §Epidemiology and Biostatistics, Memorial Sloan-Kettering Cancer Center, New York, NY Purpose: To examine the relationship between tumor regression grade (TRG) and outcomes in patients with rectal cancer treated with preoperative therapy. Methods and Materials: Specimens from 144 patients with cT3,4 rectal cancer who had received preoperative radiation ⴞ chemotherapy and had a minimum follow-up of 3 years were retrospectively reviewed. TRG, which involves examining the residual neoplastic cells and scoring the degree of both cytological changes, including nuclear pyknosis or necrosis and/or eosinophilia, as well as stromal changes, including fibrosis (either dense or edematous) with or without inflammatory infiltrate and giant-cell granulomatosis around ghost cells and keratin, was quantified in five grades according to the Mandard score (Cancer 1994;73: 2680 –2686). The greater the response, the lower the TRG score. The median follow-up was 72 months (range, 40 –143 months). Results: Of the 144 patients, 19% were TRG1, 12% were TRG2, 21% were TRG3, 46% were TRG4, and 1% were TRG5. To simplify the analysis, TRG was combined into two groups: TRG1–2 and TRG3–5. By univariate analysis, none of the pretreatment factors examined, including age, circumference, length, distance from the anorectal ring, pretreatment T and N stage, and INDpre (defined as the pretreatment reference index size based on digital rectal examination), had an impact on 5-year outcomes, including local control, metastases-free survival, disease-free survival, and overall survival. Postoperative parameters, including pathologic T stage (pT), pathologic N stage (pN), and TRG, did significantly influence 5-year outcomes. These included local failure: pT0 –2: 5% vs. pT3– 4: 19%, p ⴝ 0.007; pN0: 7% vs. pN1–3: 26%, p ⴝ 0.002; TRG1–2: 2% vs. TRG3–5: 17%, p ⴝ 0.013; metastasis-free survival: pT0 –2: 86% vs. pT3– 4: 62%, p ⴝ 0.005; pNⴚ: 86% vs. pNⴙ: 42%, p < 0.001; TRG1–2: 91% vs. TRG3–5: 66%, p ⴝ 0.004; disease-free survival: pT0 –2: 83% vs. pT3– 4: 54%, p ⴝ 0.001; pN0: 80% vs. pN1–3: 39%, p < 0.001; TRG1–2: 91% vs. TRG3–5: 58%, p < 0.001; and overall survival: pT0 –2: 85% vs. pT3– 4: 65%, p ⴝ 0.007; pN0: 86% vs. pN1–3: 45%, p < 0.001; TRG1–2: 89% vs. TRG3–5: 68%, p ⴝ 0.004. By multivariate analysis combining all pre- and posttreatment parameters, only pN (p < 0.001) and TRG (p ⴝ 0.005) significantly predicted disease-free survival. Furthermore, TRG predicted the incidence of pathologic nodal involvement (p < 0.0001). Conclusions: By univariate analysis, TRG is a predictor for local failure, metastases-free survival, and overall survival. By multivariate analysis, it predicts improved disease-free survival. Given the ability of TRG to predict those patients with Nⴙ disease, it may be helpful, in combination with other clinicopathologic factors, in selecting patients for a more conservative procedure, such as local excision rather than radical surgery, after preoperative therapy. © 2005 Elsevier Inc. Neoadjuvant chemoradiation, Prognostic factors, Rectal neoplasms.
Reprint requests to: Vincenzo Valentini, M.D., Istituto di Radiologia, Università Cattolica del Sacro Cuore, Policlinico Universitario, “Agostino Gemelli,” Largo A. Gemelli, 8, 00168 Roma, Italy. Tel: (⫹39) 6-30154376; Fax: (⫹39) 6-35511338; E-mail: [email protected]
These data were presented during the poster discussion session at the 44th Annual Meeting of ASTRO in New Orleans, LA, on October 6 –10, 2002. Received July 13, 2004, and in revised form Oct 28, 2004. Accepted for publication Nov 8, 2004. 752
TRG in rectal cancer
INTRODUCTION In the last several years, a number of studies examining the use of preoperative chemoradiation (CMT) in patients with cT3– 4 rectal cancer have reported pathologic complete response rates (pCR) of 9 –29%, as well as an increased ability to perform sphincter-sparing surgery (1– 6). Many series report that patients who achieve a pCR after preoperative CMT, independent of their initial clinical T and N stage, have improved long-term outcomes, including local control, metastases-free survival, and overall survival (6 –12). In patients who achieve less than a pCR, there is heterogeneity in definitions and techniques of identifying and scoring the presence of residual tumor after preoperative CMT. For example, stage pT3 has been defined as responses ranging from gross disease remains in the perirectal fat to those with a few foci of microscopic residual disease outside the bowel wall. Most series report that the degree of response is predictive of outcomes (6, 10). For example, Wheeler et al. have suggested that patients with only microscopic foci in the mesorectum had a better prognosis compared with patients with T3 (transmural) disease (13). The aim of this study is to evaluate the relationship between the tumor regression grade (TRG) pathologic scoring system, which is based on the characteristics of the disposition of residual neoplastic cells in the specimen, and 5-year outcomes in patients with rectal cancer treated with preoperative radiation therapy with or without concomitant chemotherapy. METHODS AND MATERIALS A total of 216 patients with cT3– 4 and/or N⫹ rectal cancer were treated with preoperative radiation with or without concurrent chemotherapy between March 1990 and February 2000 in the Department of Radiation Oncology at the Catholic University, Rome, Italy. For this analysis, 144 patients were identified who had a minimum follow-up of 3 years and the availability of the pathologic specimen for the determination of TRG score. The remaining 72 patients were excluded because of inadequate follow-up (n ⫽ 5) or the lack of pathologic material for re-review (n ⫽ 67). The pathologic specimens were reviewed in serial section and evaluated to define the TRG by two independent pathologists (V.F.M., R.R.) who were not aware of the patient’s treatment protocol or outcome. In the rare cases of disagreement, the final assessment was achieved by the same pathologists using a multihead microscope. All patients were enrolled in previously reported Phase I/II trials using a variety of preoperative radiation or CMT regimens (4, 14 –18). The radiation therapy doses ranged from 37.8 Gy to 50.4 Gy at 1.8 Gy/fraction, and the chemotherapeutic agent(s) used included one or more of the following: 5-fluorouracil (5-FU), mitomycin C, cisplatin, oxaliplatin, and raltitrexed.
Eligibility criteria and diagnostic workup Eligibility criteria for all of the protocols were the same and included the following: the most inferior aspect of the primary tumor’s being between 0 and 10 cm from the anorectal ring as evaluated by double-contrast barium enema, histologically con-
● F. M. VECCHIO et al.
753
firmed adenocarcinoma, clinical stage T3–T4 and/or N1–N3, M0 according to the 1992 American Joint Committee on Cancer staging system (19), no prior radiation therapy to the pelvis, Eastern Cooperative Oncology Group performance status of 0 –2, granulocyte count greater than 3,000 per L, platelet count greater than 100,000 per L, hemoglobin concentration greater than 10 dL, serum creatinine value not greater than 1.5 mg/dL, creatinine clearance ⬎65 mL/min, no major intercurrent disease, adequate heart function, age ⬎18 years, and informed consent. Pretreatment workup included digital rectal examination, pelvic CT scan, transrectal ultrasound, liver ultrasonography, chest Xray, double-contrast barium enema, and proctoscopy with biopsy. Four to 5 weeks after the end of preoperative treatment, preoperative restaging was performed. The clinical response was determined during a weekly meeting of all specialists involved in the diagnostic investigation. Data from individual physicians’ examinations were compared, and the definitive combined staging and the tumor response were recorded (20). In the event of a disagreement, majority consensus was used. For the cN stage, any visible nodes on CT scan were considered positive, and patients with nodes identified in the region of the inferior mesenteric artery were considered N3.
Treatment schedules The patients were treated on one of six different preoperative sequential protocols performed between 1990 and 2000. 1. T3-FUMIR: 40 patients (28%) received bolus i.v. mitomycin C (10 mg/m2, Day 1) plus continuous infusion 5-FU (1,000 mg/m2, Days 1– 4), combined with 37.8 Gy (14). 2. T3-IORT: 23 patients (16%) received 37.8 Gy, and at surgery a 10 Gy intraoperative boost (IORT) was delivered to the presacral region with 6 MeV electrons (15). 3. T3-PLAFUR: 41 patients (25%) were treated with a 1– 4-h infusion of cisplatin (60 mg/m2, Days 1, 29) plus continuous infusion 5-FU (1,000 mg/m2, Days 1– 4/5 and 29 –32/33) and concurrent 45 Gy followed by a boost to 50.4 Gy (4). 4. T3-TOMRT: 7 patients (5%) received raltitrexed (3 mg/m2 Days 1, 19, 38) plus concurrent 45 Gy followed by a boost to 50.4 Gy. At surgery, a 10 Gy IORT boost was delivered to the presacral region with 6 MeV electrons (16). 5. T3-TOMOXRT: 9 patients (6%) received raltitrexed (3 mg/ m2) and dose escalation with oxaliplatin (60 – 85–110 –130 mg/m2 Days 1, 19, 38) concurrent with 45 Gy followed by a boost to 50.4 Gy (17). 6. T4-FUMIR: 24 patients (17%) received bolus i.v. mitomycin C (10 mg/m2, Day 1) plus continuous infusion 5-FU (1,000 mg/m2 Weeks 1, 5) concurrent with 45– 48 Gy. At surgery, 4 patients received a 10 Gy IORT boost to the presacral region with 6 MeV electrons (18). Overall, 23 patients (16%) were treated with preoperative radiotherapy alone, 121 (84%) with preoperative CMT, and 29 (20%) with postoperative adjuvant chemotherapy (Table 1).
Radiation therapy techniques The clinical target volume 2 (CTV2) included the tumor, the mesorectum, and the internal iliac nodes, with or without external iliac lymph nodes. A 4-field or 3-field technique was used. The lateral border of anteroposterior-posteroanterior radiation fields was 1.5–2 cm outside the true bony pelvis; the inferior border was 1 cm above the anal verge in the tumors of the middle rectum and
754
I. J. Radiation Oncology
● Biology ● Physics
Table 1. Patient characteristics No. Age (years) Median (range) 64 (25–81) Gender Male Female cTNM stage T2N0 T2N1 T2N2 T3N0 T3N1 T3N2 T3N3 T4N0 T4N1 T4N2 T4N3 Treatment Preoperative radiotherapy alone Preoperative chemoradiation Adjuvant chemotherapy Histologic classification Adenocarcinoma G1–G2 Adenocarcinoma G3 Mucinous adenocarcinoma Adenocarcinoma NOS
%
Volume 62, Number 3, 2005
scopic residual disease. IORT was delivered using 6 MeV electrons, and the dose was 10 Gy.
Postoperative chemotherapy
96 48
66.6 33.3
1 3 1 23 52 38 3 1 6 13 3
0.7 2.1 0.7 15.9 36.1 26.3 2.1 0.7 4.1 9 2.1
23 121 29
15.9 84.1 20.1
65 13 2 63
45 9 1.4 44.4
just below the anal verge in the tumors of the lower rectum; the superior border was at least 2 cm above the tumor and not inferior to the sacral promontory; corner blocks or multileaf collimators were used to exclude extrapelvic normal tissues. The lateral field posterior border was a minimum of 1.5 cm behind the anterior bony sacral margin and the anterior border at the most posterior aspect of the symphysis pubis. In patients with T4 disease, this border was anterior to the symphysis pubis to include external iliac nodes. The clinical target volume 1 (CTV1) included the mesorectum corresponding to the tumor mass with 2-cm radial margins. Radiation was delivered with a 10 MV linear accelerator. Fractionation was at 180 cGy/day, 5 fractions per week. All fields were treated daily. The dose was prescribed to the axis intersection according to ICRU 50/62 (21, 22).
Surgery Surgery was performed 4 to 8 weeks after the completion of radiation. The choice of the surgical procedure (abdominoperineal resection [APR], low anterior resection [LAR], or local excision) was at the surgeon’s discretion. Temporary colostomy after LAR was performed at the surgeon’s discretion. For patients who underwent an LAR or APR, the resection of whole mesorectum (total mesorectal excision) was performed and inked distal and radial margins were evaluated. For those undergoing an LAR, a distal rectal margin of at least 2 cm was obtained.
Intraoperative radiation therapy A total of 34 patients with cT3 (n ⫽ 30) or cT4 (n ⫽ 4) disease were referred by a team of surgeons involved in a pilot study of IORT (20). The IORT target was chosen according to the size of the resected tumor and of the area thought to be at risk of micro-
All patients with stage cT4 (pretherapy), as well as those on protocols T3-TOMRT and T3-TOMOXRT with pN⫹ (n ⫽ 29), received postoperative chemotherapy as part of their primary treatment. The regimen was bolus 5-FU (400 mg/m2/day) and leucovorin (100 mg/m2 Days 1–5) monthly for 6 months (23).
Follow-up Patients were seen in routine follow-up every 3 months for the first year, every 6 months for the next 2 years, and then yearly. At each follow-up visit, a digital examination and proctoscopy were performed. Liver ultrasound, chest X-ray, pelvic CT, and carcinoembryonic antigen were performed at 3–12-month intervals (20). The median follow-up was 72 months (range: 40 –143 months). No patient was lost to follow-up.
Evaluation of response The tumor response was assessed according to the World Health Organization score (24). Clinical response on digital rectal examination was also evaluated using a reference index (IND) defined as the product of the quarters of circumference of rectal wall involved (categorized 1 to 4) multiplied by the craniocaudal length of the tumor in mm (4). This index was calculated at the time of initial diagnosis (INDpre) and 4 to 6 weeks after radiation (INDpost). INDpre was calculated in all patients, and INDpost was available in 97 patients. Downstaging was defined as any reduction of stage comparing clinical T (cT) and N (cN) stage to pathologic T (pT) and N (pN) stage.
Tumor regression grade Tumor regression grade (TRG) as reported by Mandard et al. in patients treated for esophageal cancer was used to assess the pathologic tumor response after preoperative therapy (25). Tumor regression after preoperative therapy was assessed by examining the residual neoplastic cells and scoring the degree of both cytological changes, including nuclear pyknosis or necrosis and/or eosinophilia, and stromal changes, including fibrosis (either dense or edematous) with or without inflammatory infiltrate and giantcell granulomatosis around ghost cells and keratin. On the basis of the combination of these changes, tumor regression was classified according to the five following grades (Fig. 1): TRG1: Complete response with absence of residual cancer and fibrosis extending through the wall. TRG2: Presence of residual cancer cells scattered through the fibrosis. TRG3: Increase in the number of residual cancer cells, with fibrosis predominant. TRG4: Residual cancer outgrowing fibrosis. TRG5: Absence of regressive changes.
Statistical analysis The distribution of ordered categories was analyzed by Gamma, Kendall’s, and Stuart’s tau tests (26). When one categorical variable was dichotomous and the other ordered, the test for linear trend of proportions was used (27). Differences of p ⱕ 0.05 were considered statistically significant. Local survival and disease-free and overall survival were estimated according to the Kaplan-Meier
TRG in rectal cancer
actuarial method. Statistical significance of each observed variable as a predictor of outcomes was analyzed by the Cox proportional hazard model (28). The clinical end points evaluated were local control, metastases-free survival, disease-free survival, and overall survival. Differences of p ⱕ 0.05 were considered statistically significant.
RESULTS Patient characteristics (pretreatment) A total of 144 patients (96 male, 48 female) with a minimum follow-up of 3 years had TRG analysis of pathologic specimen. The median age was 64 years (range, 25– 81 years). The tumor stage at initial (pretreatment) diagnosis and tumor grade are seen in Table 1. The median length of the tumor was 50 mm (SD, 17.6, range, 10 –150 mm). The number of quarters of rectal wall invaded by the tumor were measured on CT scan images; the results are as follows: 1 wall, 8 (6%); 2 walls, 69 (48%); 3 walls, 17 (12%); and 4 walls (circumferential), 50 (35%). IND was calculated as the product of number of quarters infiltrated by the tumor and the length of the tumor. The median INDpre was 120 mm (SD, 92.9, range, 25– 600 mm). The location of the tumor was defined by measuring the distance of the most inferior aspect of the tumor from the anorectal ring as seen on barium enema. The median distance was 40 mm (SD, 26.6; range, 12.5–120 mm). In 56 patients (39%) the tumor was located between 0 and 30 mm, in 43 patients (30%) it was located between 30 and 50 mm, and in 45 patients (31%) it was more than 50 mm. Response Four to 6 weeks after the completion of preoperative therapy, all patients were reevaluated with interval history and physical examination, CT scan, and barium enema. Tumor response comparing INDpre with INDpost was assessed using World Health Organization criteria. The overall clinical response rate was 51%: 4% clinical complete response and 47% partial response; the majority (49%) had a minor response that was classified as no change. No patients developed disease progression. INDpost was available in 97 patients; it ranged from 10 to 280 with a median value of 40 (SD, 54.7). Surgical procedures All patients underwent surgery after preoperative therapy; the procedures included local excision: 2 (1%), LAR: 82 (57%), LAR plus IORT: 11 (8%), APR: 29 (20%), APR plus IORT: 18 (13%), and Hartmann procedure: 2 (1%). Sphincter-sparing surgery was performed in 66% of patients. In the 56 patients in whom the most inferior aspect of the tumor was 0 –30 mm from the anorectal ring, 17 (30%) underwent sphincter-sparing surgery. Downstaging Posttreatment pathologic stages included pT0N0: 25 (17%), pT0N1: 2 (1%), pT1N0: 8 (6%), pT1N1: 1 (1%),
● F. M. VECCHIO et al.
755
pT2N0: 21 (15%), pT2N1: 3 (2%), pT2N2: 1 (1%), pT3N0: 44 (31%), pT3N1: 15 (10%), pT3N2: 12 (8%), pT3N3: 3 (2%), pT4N0: 3 (2%), and pT4N1: 6 (4%). Overall, the percentage of downstaging was 53% (77/144) for T stage 77% and 78% (93/119) for N stage. The tumor regression grade was TRG1: 27 (19%), TRG2: 18 (13%), TRG3: 31 (22%), TRG4: 67 (47%), and TRG5: 1 (1%). To simplify the analysis, TRG was combined into two groups, TRG1–2 and TRG3–5. The relationship between TRG and pN was analyzed. Overall, 41 patients (28%) with TRG1–2 were pN0, and 4 (3%) with TRG1–2 were pN1–3. In contrast, 60 patients (42%) with TRG3–5 were pN0, and 39 (27%) with TRG3–5 were pN1–3, p ⬍ 0.0001 (Table 2). Therefore, 91% (41/45) of patients with TRG1–2 were pN0, whereas 63% (60/95) of patients with TRG3–5 were pN0. Patterns of failure and survival A total of 43 patients died: 34 deaths were due to tumor, and 9 were due to intercurrent disease. Of the remaining 101 patients, 53 are without evidence of disease. Patterns of failure in the remaining 48 patients include the following: local failure: 9 (7%), distant metastasis: 29 (20%), and both local failure and distant metastasie: 10 (7%). The sites of distant metastases were as follows: liver: 22 (15%), lung: 10 (7%), brain: 2 (1%), bone: 2 (1%), and other nonspecified sites in 3 (2%). The 5-year local failure, metastasis-free survival, disease-free survival, and overall survival were 87%, 73%, 67%, and 74%, respectively. Outcome according to pretreatment prognostic factors Univariate analysis of the impact of age, tumor circumference, length and distance from the anorectal ring, cT and cN stage, and INDpre is seen in Table 3. Younger patients had a significantly higher incidence of local failure (⬍50 years: 33%, 51–70 years: 9%, and ⬎70 years: 0%, p ⫽ 0.001) and lower 5-year disease-free survival (⬍50 years: 45%, 51–70 years: 71%, and ⬎70 years: 88%, p ⫽ 0.012). Tumor circumference and length were not prognostic factors. Overall 5-year survival was significantly lower for more distal tumors (⬍30 mm: 68% vs. 31–50 mm: 89%, p ⫽ 0.029). However, there was no correlation with cT stage. Clinical nodal status significantly correlated with disease-free survival, metastasis-free survival, and overall survival. The 5-year disease-free survival by nodal status was cN0: 78%, cN1: 78%, cN2: 53%, and cN3: 50%, p ⫽ 0.049. The 5-year metastasis-free survival was cN0: 78%,
Table 2. Correlation between TRG and pN (p ⬍ 0.0001) TRG
pN0
pN1–3
TRG1–2 TRG3–5
41 (28%) 60 (42%)
4 (3%) 39 (27%)
Abbreviation: TRG ⫽ tumor regression grade.
756
I. J. Radiation Oncology
● Biology ● Physics
Volume 62, Number 3, 2005
Table 3. Univariate analysis of pretreatment parameters: Outcome at 5 years
Age Circumference Length Distance from the anorectal ring cT cN INDpre
Local failure
Metastasis-free survival
Disease-free survival
Overall survival
p ⫽ 0.001 NS NS NS NS NS NS
p ⬍ 0.05 NS NS NS NS p ⫽ 0.014 NS
p ⫽ 0.012 NS NS NS NS p ⫽ 0.049 NS
NS NS NS p ⫽ 0.029 NS p ⬍ 0.029 NS
Abbreviation: NS ⫽ nonsignificant.
cN1: 87%, cN2: 56%, and cN3: 67%, p ⫽ 0.014. The 5-year overall survival was cN0: 87%, cN1: 78%, cN2: 66%, and cN3: 50%, p ⫽ 0.029. The following 5-year outcomes decreased with an increasing number of positive nodes (cN0 –1 vs. cN2–3): local failure (9% vs. 17%, p ⫽ 0.316), metastasis-free survival (85% vs. 57%, p ⫽ 0.002), diseasefree survival (78% vs. 53%, p ⫽ 0.005), and overall survival (81% vs. 65%, p ⫽ 0.003). There was no correlation between INDpre and the 5-year local failure, metastasis-free survival, disease-free survival, and overall survival. By multivariate analysis, the following variables were independent factors for outcomes. Local failure: age (p ⫽ 0.006) and distance of tumor from the anorectal ring (p ⫽ 0.033); metastasis-free survival: distance of tumor from anorectal ring (p ⫽ 0.003), length of the tumor (p ⫽ 0.015), lymph node status (p ⫽ 0.045), and INDpre (p ⫽ 0.019); disease-free survival: age (p ⫽ 0.019), distance of tumor from the anorectal ring (p ⫽ 0.003), lymph node status (p ⫽ 0.040), and INDpre (p ⫽ 0.029); and overall survival: distance of tumor from the anorectal ring (p ⫽ 0.004) and lymph node status (p ⫽ 0.016). Outcomes according to posttreatment prognostic factors As seen in Table 4, by univariate analysis, except for the distance from the anorectal ring, all posttreatment factors (pT, pN, INDpost, TRG) significantly predicted outcomes. Compared with patients with INDpost ⬍30, those with INDpost ⬎30 had a significant decrease in 5-year diseasefree survival (64% vs. 76%, p ⫽ 0.025) and overall survival (67% vs. 86%, p ⫽ 0.043). There was a significant decrease in favorable outcomes with increasing pT stage. This included 5-year local failure (pT0 –1: 4% vs. pT3– 4: 19%, p ⫽ 0.007), metastasis-free survival (pT0 –1: 86%
vs. pT3– 4: 62%, p ⫽ 0.005), disease-free survival (pT0 –1: 83% vs. pT3– 4: 54%, p ⫽ 0.001), and overall survival (pT0 –1: 85% vs. pT3– 4: 65%, p ⫽ 0.007). Likewise, there was a significant decrease in outcomes with increasing pN stage. This included 5-year local failure (pN0: 7% vs. pN1–3: 26%, p ⫽ 0.002), metastasis-free survival (pN0: 86% vs. pN1–3: 42%, p ⬍ 0.001), 5-year disease-free survival (pN0: 80% vs. pN1–3: 40%, p ⬍ 0.001), and overall survival (pN0: 86% vs. pN1–3: 45%, p ⬍ 0.001). With the exception of local control, as TRG increased, the corresponding 5-year outcomes decreased. The cumulative rate of distant metastases by TRG was TRG1: 93%, TRG2: 90%, TRG3: 86%, and TRG4: 56% (p ⫽ 0.001). Similar results were seen for 5-year disease-free survival; TRG1: 96%, TRG2: 90%, TRG3: 74%, and TRG4: 52% (p ⫽ 0.002). The overall rate of death was TRG1: 11%, TRG2: 5%, TRG3:19%, and TRG4: 42% (p ⫽ 0.002); 5-year survival was TRG1: 91%, TRG2: 85%, TRG3: 79%, and TRG4: 63% (p ⫽ 0.016). When TRG was combined into two groups (TRG1–2 vs. TRG3–5), the differences in outcomes remained significant (Fig. 1). This included 5-year local failure: TRG1–2: 2% vs. TRG3–5: 17%, p ⫽ 0.013 (Fig. 2); metastasis-free survival: TRG1–2: 91% vs. TRG3–5: 66%, p ⫽ 0.004 (Fig. 3); disease-free survival: TRG1–2: 91% vs. TRG3–5: 58%, p ⬍ 0.001 (Fig. 4); and overall survival: TRG1–2: 89% vs. TRG3–5: 68%, p ⫽ 0.004 (Fig. 5). By multivariate analysis, a number of factors significantly predicted 5-year outcomes. These included local failure: pN (p ⫽ 0.020) and TRG (p ⫽ 0.035); metastasis-free survival: pT (p ⫽ 0.010), pN (p ⬍ 0.001), and TRG (p ⫽ 0.005); disease-
Table 4. Univariate analysis of posttreatment parameters: Outcomes at 5 years
pT pN TRG1–2 vs. TRG3–5 INDpost Abbreviation: NS ⫽ nonsignifant.
Local failure
Metastasis-free survival
p ⫽ 0.007 p ⫽ 0.002 p ⫽ 0.013 NS
p ⫽ 0.005 p ⬍ 0.001 p ⫽ 0.004 NS
Disease-free survival p p p p
⫽ ⬍ ⬍ ⫽
0.001 0.001 0.001 0.025
Overall survival p p p p
⫽ ⬍ ⫽ ⫽
0.007 0.001 0.004 0.043
TRG in rectal cancer
Fig. 1. Tumor regression grade (TRG).
free survival: pN (p ⬍ 0.001) and TRG (p ⫽ 0.003); and overall survival: pN (p ⬍ 0.001) and TRG (p ⫽ 0.042). Combined multivariate analysis of both pre- and posttreatment factors Multivariate analysis of all pre- and posttreatment factors revealed that local failure, pN (p ⫽ 0.024), and age (p ⫽ 0.025) were independent prognostic factors. These included metastasis-free survival: pT (p ⫽ 0.037), pN (p ⬍ 0.001), and TRG (p ⫽ 0.005); disease-free survival: pN (p ⬍ 0.001) and TRG (p ⫽ 0.011); overall survival: pN (p ⬍ 0.001).
● F. M. VECCHIO et al.
757
Morgan et al. tested the interobserver reliability using TRG to assess pathologic tumor response after preoperative CMT in 21 patients with T3–T4 rectal cancer. A TRG score of 1–3 was seen in 62%, and TRG4 –5 was seen in 38%. The pCR rate was 14%. In contrast to our results, they observed that pT and pN stages were not directly related to TRG. However, the interobserver reliability in scoring TRG was high. The authors did not analyze the impact of TRG grade on outcomes (29). Other grading systems to quantify tumor regression after preoperative treatment have been developed. Wheeler et al. retrospectively reviewed surgical specimens of 42 patients with T3/T4 rectal cancer treated with preoperative CMT (13). The specimens were analyzed by one pathologist, and the Rectal Cancer Regression Grade (RCRG) was measured. The authors defined three grades of regression: RCRG 1: either pCR or only microscopic foci of adenocarcinoma, RCRG 2: marked fibrosis with macroscopic tumor still present, and RCRG 3: a poor response with little or absence of fibrosis in the presence of abundant macroscopic tumor. Overall, 46% were RCRG 1, 40% were RCRG 2, and 14% were RCRG 3. The authors speculated that RCRG better defines regression of disease after CMT compared with the pT or pN stage. The subset of patients with pT3 microscopic foci had a more favorable prognosis compared with those with pT3 macroscopic disease. At Memorial Sloan-Kettering Cancer Center, selected clinicopathologic factors, including the extent of pathologic response to preoperative radiation or CMT, were retrospec-
DISCUSSION Retrospective data suggest that pathologic downstaging after preoperative radiotherapy with or without concomitant chemotherapy is associated with improved outcomes (6 –10, 12). Patients with pT0 –2 stage disease after preoperative therapy have local failure rates of 0 – 6% and 5-year survivals of 90 –100%. However, there is heterogeneity in the definitions as well as the techniques of identifying and scoring the presence of residual tumor after preoperative therapy. Therefore, we have analyzed response using the TRG score, which offers a more uniform method of assessment. The TRG score was developed by Mandard et al. to assess the response of preoperative therapy in patients with esophageal cancer. They examined 93 resected specimens of esophageal cancer from patients treated with preoperative CMT. TRG score predicted 3-year survival independent of tumor stage; TRG1: 61%, TRG2: 60%, TRG3: 27%, TRG4: 0%, and TRG5: 0%, p ⬍ 0.0001. By univariate analysis, TRG, pN stage, tumor size, and the amount of wall involvement by tumor were significantly correlated with diseasefree survival (p ⬍ 0.05). By multivariate analysis, only TRG (TRG1–3 vs. TRG4 –5) remained a significant predictor of disease-free survival (p ⬍ 0.001). The authors recommended that TRG should be considered when evaluating therapeutic results after preoperative CMT in esophageal cancer (25).
Fig. 2. Local control according to TRG1–2 vs. TRG3–5 (p ⫽ 0.013). TRG ⫽ tumor regression grade.
758
I. J. Radiation Oncology
● Biology ● Physics
Fig. 3. Metastases-free survival according to TRG1–2 vs. TRG3–5 (p ⫽ 0.004). TRG ⫽ Tumor regression grade.
tively analyzed to examine the impact on long-term recurrence-free survival in 69 patients with T3– 4 and/or pN⫹ primary rectal cancer. A 5-point grading scale was used to categorize pathologic response to preoperative therapy: 0: no response, 1: response ⱕ1/3, 2: response 1/3–2/3, 3: response ⱖ2/3 but ⬍ complete, and 4: complete response. With a median follow-up of 69 months, 5-year relapse-free survival was 79% and was significantly worse for patients with aggressive pathologic features and pN⫹. Risk ratios for relapse-free survival were 3.68 for the presence of aggressive pathologic features and 4.64 for node-positive rectal cancers. In patients with greater than 95% response to preoperative therapy, only 1 patient was dead from disease; another died of an unrelated cause, and the remaining patients were free of disease with a minimum follow-up of 47 months. However, the extent of response did not predict relapse-free survival (12). Rodel et al. examined the apoptotic index (AI), Ki-67, p-53, and p-2 immunohistochemically on pretreatment biopsies in 44 patients with T3/4 rectal cancer after preoperative CMT (30). The treatment response was assessed with a 5-grade system using histopathologic methods on the resected surgical specimen. The AI was obtained by dividing the number of apoptotic tumor cells by the total number of tumor cells multiplied by 100; the pathologic response of the tumor was defined according to a rectal cancer regression grading defined by Dworak et al. (31). This includes Grade 0: no regression, Grade 1: minimal regression, Grade 2: moderate regression, Grade 3: good regression,
Volume 62, Number 3, 2005
Grade 4: total regression. Tumors with Grades 3 or 4 response to CMT had significantly higher pretreatment levels of apoptosis than tumors with lower regression (Grades 0 –2). The pretreatment AI predicted the risk of 5-year local failure: 100% for AI greater than median vs. 74% for AI less than median (p ⫽ 0.08). Furthermore, AI predicted tumor regression. Tumors with Grades 3/4 response to CMT had significantly higher pretreatment levels of apoptosis (mean AI: 2.06%) than tumors with Grades 0 –2 response (AI: 1.44%, p ⫽ 0.003). A significant correlation was also found between pathologic tumor response and relapse-free survival (Grade 3/4: 72% vs. Grade 0/1/2: 28%, p ⫽ 0.004). When the analysis was limited to local control, the only pretreatment factor significantly correlated with this end point was AI. Moreover, AI significantly correlated with Ki-67 but not with p-53 and bcl2 status. The authors concluded that AI and tumor regression can be helpful in selecting preoperative therapy (30). In our experience, by univariate analysis, pretreatment parameters (age, circumference, length and distance from the anorectal ring, pretreatment T and N stage, and INDpre) were associated with some but not all 5-year outcomes. Posttreatment parameters (pT, pN, and TRG) affected all 5-year outcomes. By multivariate analysis, pooling all preand posttreatment parameters, only pN and TRG significantly predicted disease-free survival. Furthermore, TRG significantly predicted the presence of pN⫹ disease. We have previously reported the impact of pathologic staging and of tumor index after treatment (INDpost) in 165
Fig. 4. Disease-free survival according to TRG1–2 vs. TRG3–5 (p ⬍ 0.001). TRG ⫽ tumor regression grade.
TRG in rectal cancer
● F. M. VECCHIO et al.
759
analysis, INDpost ⬍30 (i.e., ⱕ1/4 circumference ⱕ30 mm length) was predictive of 5-year local control and survival. Our data support the role of the TRG score in predicting 5-year outcomes. In contrast to INDpost, TRG offers a more reliable interpretation of the tumor regression in the pathologic specimen, accounting for the presence of residual isolated cells in a well-defined group at a lower risk. Not only does it decrease the chance of misinterpretation of the pathologic T stage as described by Morgan et al. (29), but the interobserver analysis is also more reliable. It should be emphasized that in our prior series examining INDpost, patients with residual isolated cancer cells in the specimen were scored as pT1, pT2, or pT3 based on the location of residual cells in the rectal wall or in the mesorectum (10). As with other clinicopathologic scoring systems, our preliminary, retrospective data need to be validated in a larger, prospective trial. It must be emphasized that it is retrospective with unknown selection bias related to the exclusion of patients with lack of pathologic material for re-review, and variables such as the impact of chemotherapy, radiation doses, the interval between radiation and surgery, and the pretreatment stage may account for the differences in results reported. Fig. 5. Overall survival according to TRG1–2 vs. TRG3–5 (p ⫽ 0.004). TRG ⫽ tumor regression grade.
CONCLUSION
consecutive patients with locally advanced rectal cancer, treated with preoperative 5-FU– based CMT (4). Increasing pT stage was a significant predictor for 5-year survival (pT0: 91%, pT2: 81%, pT3: 66%, and pT4: 47% [p ⫽ 0.014]), as well as local failure (pT0: 0%, pT1: 17%, pT2: 12%, pT3: 21%, and pT4: 54% [p ⫽ 0.0012]). The pN status significantly predicted freedom from distant metastases and overall survival. By univariate and multivariate
Tumor regression grade seems to be a prognostic factor for disease-free survival in patients receiving preoperative therapy for rectal cancer. In addition, it is a prognostic factor for local failure, metastasis-free survival, and overall survival. Given the ability of TRG to predict those patients with N⫹ disease, in combination with other clinicopathologic factors, it may be helpful after preoperative therapy in selecting patients for a more conservative procedure, such as local excision rather than radical surgery.
REFERENCES 1. Rich TA, Skibber JM, Ajani JA, et al. Preoperative infusional chemoradiation therapy for stage T3 rectal cancer. Int J Radiat Oncol Biol Phys 1995;32:1025–1029. 2. Chari RS, Tyler DS, Anscher MS, et al. Preoperative radiation and chemotherapy in the treatment of adenocarcinoma of the rectum. Ann Surg 1995;221:778 –786. 3. Grann A, Minsky BD, Cohen AM, et al. Preliminary results of preoperative 5-fluorouracil, low-dose Leucovorin and concurrent radiation therapy for clinically resectable T3 rectal cancer. Dis Colon Rectum 1997;40:515–522. 4. Valentini V, Coco C, Cellini N, et al. Ten years of preoperative chemoradiation for extraperitoneal T3 rectal cancer: Acute toxicity, tumor response, and sphincter preservation in three consecutive studies. Int J Radiat Oncol Biol Phys 2001; 51:371–383. 5. Bosset JF, Magnin V, Maingon P, et al. Preoperative radiochemotherapy in rectal cancer: Long-term results of a Phase II trial. Int J Radiat Oncol Biol Phys 2000;46:323–327. 6. Mohiuddin M, Hayne M, Regine WF, et al. Prognostic significance of postchemoradiation stage following preoperative chemotherapy and radiation for advanced/recurrent
7.
8.
9.
10.
11.
rectal cancers. Int J Radiat Oncol Biol Phys 2000;48: 1075–1080. Janjan NA, Abbruzzese J, Pazdur R, et al. Prognostic implications of response to preoperative infusional chemoradiation in locally advanced rectal cancer. Radiother Oncol 1999;51: 153–160. Berger C, de Muret A, Garaud P, et al. Preoperative radiotherapy (RT) for rectal cancer: Predictive factors of tumor downstaging and residual tumor cell density (RTCD): Prognostic implications. Int J Radiat Oncol Biol Phys 1997;37: 619 – 627. Kaminsky-Forrett MC, Conroy T, Luporsi E, et al. Prognostic implications of downstaging following preoperative radiation therapy for operable T3-T4 rectal cancer. Int J Radiat Oncol Biol Phys 1998;42:935–941. Valentini V, Coco C, Picciocchi A, et al. Does downstaging predict improved outcome after preoperative chemoradiation for extraperitoneal locally advanced rectal cancer? A longterm analysis of 165 patients. Int J Radiat Oncol Biol Phys 2002;53:664 – 674. Gerard JP, Chapet O, Morignat E, et al. Preoperative radio-
760
12. 13. 14.
15.
16.
17.
18. 19. 20.
I. J. Radiation Oncology
● Biology ● Physics
therapy of rectal cancer. The Lyons experience, 1985–1996. Prognostic study apropos of 312 patients. Ann Chir 1999;53:1003–1010. 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. Wheeler JM, Warren BF, Mortensen NJMcC, et al. Quantification of histologic regression of rectal cancer after irradiation. Dis Colon Rectum 2002;45:1051–1056. Valentini V, Coco C, Cellini N, et al. Preoperative chemoradiation for extraperitoneal T3 rectal cancer: Acute toxicity, tumor response, sphincter preservation. Int J Radiat Oncol Biol Phys 1998;40:1067–1075. Pacelli F, Di Giorgio A, Papa V, et al. Preoperative radiotherapy combined with intraoperative radiotherapy improves results of total mesorectal excision in patients with T3 rectal cancer. Dis Colon Rectum 2004;47:170 –179. Valentini V, Doglietto GB, Morganti AG, et al. Preoperative chemoradiation with raltitrexed (‘Tomudex’) for T2/N ⫹ and T3/N ⫹ rectal cancers: A phase I study. Eur J Cancer 2001; 37:2050 –2055. Valentini V, Morganti AG, Luzi S, et al. Chemoradiation with raltitrexed (‘Tomudex’)and oxaliplatin in pre-operative treatment of stage II/III resectable rectal cancer: A dose finding study. Proc ASCO 2001;20:131a. Sofo L, Ratto C, Doglietto GB, et al. Intraoperative radiation therapy in integrated treatment of rectal cancers. Results of phase II study. Dis Colon Rectum 1996;39:1396 –1403. UICC International Union Against Cancer. TNM classification of malignant tumours. 4th ed., 2nd rev. Springer Verlag; 1992. Marano P, Barbaro B, De Franco A, et al. Diagnostic and therapeutic integration in rectal cancer. Preliminary notes. Radiol Med (Torino) 1992;84:261–273.
Volume 62, Number 3, 2005
21. International Commission on Radiation Units and Measurements. ICRU Report 50: Prescribing, recording and reporting photon beam therapy. 1993. 22. International Commission on Radiation Units and Measurements. ICRU Report 62: Prescribing, recording and reporting photon beam therapy (supplement to ICRU Report 50). ICRU News, December 1999. 23. Machover D, Schwarzenberg L, Goldschmidt E, et al. Treatment of advanced colorectal and gastric adenocarcinomas with 5-FU combined with high-dose folinic acid: A pilot study. Cancer Treat Rep 1982;66:1803–1807. 24. Miller AB. Reporting result of cancer treatment. Cancer 1981; 47:207–214. 25. Mandard AM, Dalibard F, Mandard JC, et al. Pathologic assessment of tumor regression grade after preoperative chemoradiotherapy of esophageal carcinoma. Cancer 1994;73: 2680 –2686. 26. Brown MB, Benedetti JK. Sampling behavior of tests for correlation in two-way contingency tables. J Am Stat Assoc 1977;72:309 –315. 27. Cochran WG. Some methods for strengthening the common chi2 tests. Biometrics 1954;10:417– 441. 28. Cox DR. Regression model and life tables. J R Stat Soc Series B 1972;34:187–220. 29. Morgan MJ, Koorey DJ, Painter D, et al. Histological tumour response to pre-operative combined modality therapy in locally advanced rectal cancer. Colorectal Dis 2002;4:177–183. 30. Rodel C, Grabenbauer GG, Papadopoulos T, et al. Apoptosis as cellular predictor for histopathologic response to neoadjuvant radiochemotherapy in patients with rectal cancer. Int J Radiat Oncol Biol Phys 2002;52:294 –303. 31. Dworak O, Keilholz L, Hoffmann A. Pathological features of rectal cancer after preoperative radiochemotherapy. Int J Colorectal Dis 1997;12:19 –23.