- Email: [email protected]
Higher serum C-reactive protein concentration and hypoalbuminemia are poor prognostic indicators in patients with esophageal cancer undergoing radiotherapy
Radiotherapy and Oncology 92 (2009) 270–275
Contents lists available at ScienceDirect
Radiotherapy and Oncology journal homepage: www.thegreenjournal.com
Radiotherapy of esophagus cancer
Higher serum C-reactive protein concentration and hypoalbuminemia are poor prognostic indicators in patients with esophageal cancer undergoing radiotherapy Chang-Yu Wang a,b,c, Ming-Jang Hsieh d, Yi-Chun Chiu e, Shau-Hsuan Li f,*, Hurng-Wern Huang b, Fu-Min Fang c, Yu-Jie Huang c a
National Kaohsiung University of Applied Sciences, Taiwan Institute of Biomedical Sciences, National Sun Yat-Sen University, Taiwan Department of Radiation Oncology, Chang Gung Memorial Hospital-Kaohsiung Medical Center, Chang Gung University College of Medicine, Kaohsiung, Taiwan d Department of Thoracic & Cardiovascular Surgery, Chang Gung Memorial Hospital-Kaohsiung Medical Center, Chang Gung University College of Medicine, Kaohsiung, Taiwan e Division of Hepato-Gastroenterology, Chang Gung Memorial Hospital-Kaohsiung Medical Center, Chang Gung University College of Medicine, Kaohsiung, Taiwan f Department of Hematology-Oncology, Chang Gung Memorial Hospital-Kaohsiung Medical Center, Chang Gung University College of Medicine, Kaohsiung, Taiwan b c
a r t i c l e
i n f o
Article history: Received 16 February 2008 Received in revised form 12 November 2008 Accepted 5 January 2009 Available online 3 February 2009 Keywords: Esophageal cancer Radiotherapy C-reactive protein (CRP) Albumin Glasgow Prognostic score (GPS)
a b s t r a c t Background and purpose: We evaluate if C-reactive protein (CRP) is an objective biomarker of esophageal cancer in patients undergoing radiotherapy. Materials and methods: Between November 2002 and July 2007, 123 patients undergoing radiotherapy for newly diagnosed esophageal cancer were enrolled. Serum CRP concentration was measured before the initiation of treatment. The relationship between serum CRP levels and other relevant variables such as body mass index, white blood cell count, platelet count, bilirubin, aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase (ALP), cholesterol, hemoglobin, and albumin levels was also analyzed. Results: Eighty-one patients of the 123 patients enrolled (65.9%) had elevated CRP levels (P5 mg/L). The 2-year survival for patients with CRP P5 mg/L was 7.8% compared to 78.4% for patients with CRP <5 mg/ L. Hypoalbuminemia (albumin <3.5 g/dL) was also related to shorter survival using univariate analysis. Multivariate analysis demonstrated that only higher serum CRP concentration and hypoalbuminemia were independent prognostic indicators for survival of patients with esophageal cancer. Conclusions: Pretreatment serum levels of CRP and albumin are objective, easily measurable biomarkers which can be used in combination with conventional staging to accurately predict survival in patients with esophageal cancer treated with radiotherapy. Ó 2009 Elsevier Ireland Ltd. All rights reserved. Radiotherapy and Oncology 92 (2009) 270–275
Esophageal cancer is the sixth most common cause of cancer death among men in Asian countries. Ninety percent of all esophageal cancer in Asian men is squamous cell carcinoma (SCC), which is in contrast to the percentages found in men from western countries [1,2]. Esophageal cancer is usually diagnosed at an advanced stage, making curative surgical resection feasible for only 30–40% of patients [3]. Therefore, neoadjuvant concurrent chemoradiotherapy (CCRT) plays an important role in managing patients with unresectable disease at the time of initial diagnosis. Trimodality treatment provides significant improvement in outcome, which is attributable to the combination of neoadjuvant CCRT with surgery for patients with good performance status. For patients who can-
* Corresponding author. Address: Esophageal Cancer Combined Clinic, Department of Radiation Oncology, Chang Gung Memorial Hospital-Kaohsiung Medical Center, No. 123, Dapi Rd., Niaosong Township, Kaohsiung County 833, Taiwan. E-mail address: [email protected] (S.-H. Li). 0167-8140/$ - see front matter Ó 2009 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.radonc.2009.01.002
not tolerate the surgical procedure, CCRT is provided as an alternative [4]. Since the optimal treatment strategy for advanced esophageal cancer is not well established, there is a need for prognostic markers that can improve patient selection for the various treatment strategies. If surgical resection (trimodality treatment) is not planned, the break in the radiotherapy treatment regimen that is usually required for preoperative surgical evaluation can be avoided, allowing for the completion of CCRT as soon as possible. Systemic inflammation has been associated with cancers, and many patients have an acute phase protein response that can manifest as an elevation in serum C-reactive protein (CRP) concentration [5–7]. There is increasing evidence that an elevated serum CRP level before surgery has prognostic value in patients with resectable esophageal cancer [8–10]. CRP is also a prognostic factor in patients with hepatocellular carcinoma, breast cancer, epithelial ovarian cancer, and malignant fibrous histiocytoma [11–14]. There
C.-Y. Wang et al. / Radiotherapy and Oncology 92 (2009) 270–275
had been much interest in establishing CRP-based prognostic scoring systems for patients with cancer. The Glasgow Prognostic score (GPS), which is based on a combination of elevated CRP and hypoalbuminemia levels, is useful in predicting the survival of patients with non-small-cell lung cancer, breast cancer, colorectal cancer, pancreatic cancer, renal cancer, and gastro-esophageal cancer [15–20]. The purpose of our study is to identify similar serum biomarkers for esophageal cancer. We measure the pretreatment serum levels of CRP as well as the relationship between serum CRP levels and other relevant variables such as body mass index (BMI), white blood cell count, platelet count, bilirubin, aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase (ALP), total cholesterol, hemoglobin, and albumin levels. We then analyze their value in predicting the survival of patients with esophageal cancer who have undergone radiotherapy. Materials and methods Patient population and measurement of CRP Between November 2002 and July 2007, patients with newly diagnosed esophageal SCC or adenocarcinoma undergoing radiotherapy were included in our study. This study was approved by the Institutional Review Board of Chang Gung Memorial Hospital, and informed consent was obtained according to established guidelines. Serum concentration of CRP was measured using the immunonephelometry method on BN II nephelometer (Dade Behring, Marburg, Germany) before the initiation of radiotherapy. Our study used a threshold value of 5 mg/L for serum CRP as suggested by the assay manufacturer. Serum levels less than 5 mg/L were considered normal in this assay. Similar studies from Japan showed the prognostic value of an elevated CRP concentration based on the same threshold for CRP in patients undergoing surgical resection for esophageal SCC [8–10]. However, the studies of gastrointestinal cancer from the UK used a higher CRP threshold (>10 mg/L) than that used in our study [11,17]. Diverse assays in these studies accounts for the use of different CRP cutoffs. The best cutoff values for elevated serum CRP levels (>5 vs. >10 mg/L) were assessed in our study using multivariate survival analysis. In addition, the measurements of serum CRP were repeated after one week if there were initial signs of infection including fever (>38 °C) or WBC count greater than 12,000/mm3. The lowest CRP level was then used for analysis. Pretreatment evaluation Patients were evaluated by a multidisciplinary team including a thoracic surgeon, a medical oncologist, a radiation oncologist, and a gastroenterologist. Patients were staged based on the chest radiograph, barium esophagram, panendoscopy, computed tomography (CT) scan of the chest and abdomen, endoscopic ultrasound (EUS), hepatobiliary ultrasound, and bone scan. Bronchoscopy was performed whenever tracheoesophageal fistula was suspected. Positron emission tomography (PET) was performed if occult distant metastases were suspected. The clinical TNM stage was determined according to the 2002 American Joint Committee on Cancer staging system. Treatment protocol Patients classified as having unresectable lesions were treated with two cycles of cisplatin and 5-fluorouracil (5-FU)-based chemotherapy and radiotherapy, concurrently. Each cycle lasted between 4 and 6 weeks. However, patients with poor performance status did not receive chemotherapy, and were treated with radiotherapy only.
271
Three-dimensional conformal radiotherapy (3DCRT) via a fourfield technique was used in most cases. The gross tumor volume (GTV) was defined based on the CT scan and esophagram, or PET scan, if available. The clinical target volume (CTV) covered GTV with 3 cm superior and inferior margins, and 1 cm radial margins. The planning target volume (PTV) was generated from the expansion of CTV by a 1 cm margin. For upper third and lower third tumor sites, the supraclavicular lymph nodes and celiac lymph nodes were also included for prophylactic irradiation. The prescribed dose was administered to the 95% isodose line, covering PTV for 3600 cGy in 20 fractions. The maximum dose to the spinal cord was limited to 4500 cGy. No more than 40% of the normal lung tissue received a radiation dose of 2000 cGy. Response assessment before esophagectomy Within 1–2 weeks following the end of neoadjuvant CCRT, CT scan and EUS were performed to assess clinical response. The multidisciplinary team reviewed the clinical information to determine if the lesions were resectable. If the lesions were classified as resectable, a trimodality treatment was advised by performing esophagectomy approximately 4–6 weeks after the end of neoadjuvant CCRT. Pathologic complete response (pCR) was defined as the complete disappearance of all viable cancer cells in all surgical specimens including the primary esophageal tumor and lymph nodes. Patients who did not achieve a pathologic complete response received a radiotherapy boost to the tumor bed of an additional 3060 cGy in 17 fractions along with two cycles of cisplatin and 5-FU-based chemotherapy. Patients categorized as unresectable, or who refused esophagectomy after neoadjuvant CCRT, received a further radiotherapy boost up to a total dose of 6660 cGy and another two to four cycles of chemotherapy concurrently to complete the course of CCRT. Statistical analysis Continuous data were presented as mean ± standard deviation (SD). The chi-square test, Student’s t test, and Mann–Whitney U test were used to compare data between the two groups. Survival time was measured from the start of radiotherapy until death or the last date of follow-up. The Kaplan–Meier method was used for univariate survival analysis, and the difference between survival curves was tested by a log-rank test. The Cox regression model was used for the multivariate survival analysis of the clinicopathological variables. Furthermore, the receiver operating characteristic (ROC) curve was employed to determine the positive predictive value (PPV) and negative predictive value (NPV) in patients who survived more than 12 months.
Results Patient characteristics One hundred and twenty males and 3 females, whose ages ranged from 34 to 81 years (median age of 54 years at diagnosis), were enrolled in our study. The clinicopathological characteristics of the patients are shown in Table 1. Histologically, 116 (94.3%) of the patients’ tumors were classified as SCC and the remaining 7 (5.7%) were adenocarcinoma. Tumors were located in the upper third of the esophagus in 19 (15.4%) patients; in the middle third of the esophagus in 65 (52.8%), and in the lower third in 39 (31.7%). According to the TNM classification, six patients were classified as stage I, (4.9%), 16 were classified as stage II (13.0%), 57 were
272
CRP as a prognostic factor for esophageal cancer
Table 1 Relationship between serum CRP levels and clinical characteristics. Mean CRP level mg/L
p
CRP <5 mg/L
CRP P5 mg/L
Patients (n = 42)
Patients (n = 81)
p
Sex Male Female
41.0 ± 55.3 38.7 ± 53.7
0.730
42 (100.0%) 0 (0%)
78 (96.3%) 3 (3.7%)
0.207
Age <60 P60
41.5 ± 50.9 39.9 ± 63.2
0.828
29 (69.0%) 13 (31.0%)
53 (65.4%) 28 (34.6%)
0.687
Histology SCC Adenocarcinoma
38.8 ± 54.4 77.4 ± 56.1
0.025
42 (100.0%) 0 (0.0%)
74 (91.4%) 7 (8.6%)
0.050
Tumor location Upper third Middle third Lower third
67.1 ± 67.1 34.7 ± 51.4 36.2 ± 50.8
0.042
4 (9.5%) 21 (50.0%) 17 (40.5%)
15 (18.5%) 44 (54.3%) 22 (27.2%)
0.213
T stage T1-2 T3-4
38.6 ± 48.5 41.3 ± 56.0
0.572
6 (14.3%) 36 (85.7%)
8 (9.9%) 73 (90.1%)
0.465
N stage N0 N1
30.9 ± 48.7 43.5 ± 56.5
0.034
13 (31.0%) 29 (69.0%)
12 (14.8%) 69 (85.2%)
0.035
M stage M0 M1
34.1 ± 56.6 53.3 ± 50.5
<0.001
37 (88.1%) 5 (11.9%)
42 (51.9%) 39 (48.1%)
<0.001
TNM stage I–II III–IV
27.4 ± 48.2 43.9 ± 56.2
0.005
14 (33.3%) 28 (66.7%)
8 (9.9%) 73 (90.1%)
0.001
Treatment Trimodality Definitive CCRT RT only
20.4 ± 33.1 49.2 ± 57.6 64.5 ± 81.0
0.003
21 (50.0%) 19 (45.2%) 2 (4.8%)
20 (24.7%) 52 (64.2%) 9 (11.1%)
0.016
Continuous data were expressed as mean ± standard deviation. The chi-square test, Student’s t test, and Mann–Whitney U test were used to compare the data between groups. CRP: C-reactive protein; SCC: squamous cell carcinoma; CCRT: concurrent chemoradiotherapy; RT: radiotherapy.
classified as stage III (46.3%), and 44 were as stage IV (35.8%). Those patients with stages I and II disease were considered unsuitable for surgery because of comorbidity. Forty-one (33.3%) of the 123 patients underwent trimodality treatment, 71 (57.7%) patients received definitive CCRT, and the remaining 11 (8.9%) patients were treated with RT only. Of the 41 patients receiving trimodality treatment, 22 (53.7%) also underwent postoperative adjuvant CCRT. The median duration of followup for patients who were alive at the time of the most recent follow-up was 23 months (with a range of 3–57 months). Eight (6.5%) of the 123 patients died of tracheoesophageal fistula with severe pneumonia or massive hemorrhage during the course of neoadjuvant CCRT. Overall, 115 (93.5%) patients completed the planned treatment protocol.
Correlation between CRP levels and clinicopathological parameters Patients were classified as normal CRP (NC) group (n = 42) if their CRP serum level <5 mg/L. If their CRP serum level P5 mg/L, they were classified as the higher CRP (HC) group (n = 81). Patients in the HC group had significantly higher mean CRP levels than those in NC group (60.68 ± 58.92 mg/L vs. 2.98 ± 0.68 mg/L, p < 0.001). Patients with more advanced disease including lymph node metastases (N1), distant metastases (M1), and stages III–IV had significantly higher mean CRP levels than those with early stage cancer (Table 1). There was a good correlation between the treatment method (trimodality vs. CCRT vs. RT only) and the mean CRP level (p = 0.003). In addition, patients with the upper third esophageal cancer tended to have higher mean CRP levels than
Table 2 Relationship between levels of serum CRP and relevant continuous variables.
White blood cell (WBC, count/lL) Hemoglobin (g/dL) Platelet (count/lL) Lymphocyte (%) Albumin (g/dL) Aspartate aminotransferase (AST, U/L) Alanine aminotransferase (ALT, U/L) Alkaline phosphatase (ALP, U/L) Bilirubin total (mg/dL) Total cholesterol (mg/dL) Body mass index (BMI)
CRP <5 mg/L
CRP P5 mg/L
p
7324 ± 2177 13.5 ± 1.7 257690 ± 67942 25.5 ± 8.9 3.8 ± 0.6 26.6 ± 17.5 23.7 ± 40.1 80.3 ± 26.1 0.6 ± 0.3 196.9 ± 46.1 21.0 ± 2.8
8141 ± 2686 12.3 ± 2.1 263988 ± 118904 19.7 ± 10.5 3.2 ± 0.7 31.7 ± 21.3 24.6 ± 20.1 92.7 ± 36.2 0.9 ± 0.7 180.9 ± 47.4 21.0 ± 4.0
0.092 0.001 0.711 0.003 <0.001 0.189 0.870 0.056 0.034 0.140 0.985
Continuous data were presented as mean ± standard deviation. Student’s t test and Mann–Whitney U test were used to compare the data between groups. CRP: C-reactive protein.
C.-Y. Wang et al. / Radiotherapy and Oncology 92 (2009) 270–275
273
those with the middle and the lower third esophageal cancer (p = 0.042). The relationship between serum CRP levels and clinicopathological variables was analyzed (n = 123, Table 1). The proportion of lymph node metastases (N1) in the HC group was significantly higher than the NC group or 85.2% (69 of 81) compared to 69.0% (29 of 42; p = 0.035). Seventy-three of the 101 patients (72.3%) with stages III–IV disease had serum CRP levels P5 mg/L, whereas only 8 of the 22 patients (36.4%) with stages I–II disease had CRP levels P5 mg/L (p = 0.001). Thirty-seven of the 42 patients (88.1%) without distant metastases (M0) had normal CRP levels, compared to only 42 of the 81 patients (51.9%) with distant metastases (M1) (p < 0.001). No significant difference in sex, age, histological type, tumor location, or T stage was found between the HC and the NC group. The relationship between serum CRP levels and relevant continuous variables is shown in Table 2. There was no significant difference in BMI, white blood cell counts, platelet counts, AST, ALT, ALP, and total cholesterol levels between NC and HC groups. Patients in the HC group had lower lymphocyte percentages, hemoglobin, and albumin levels compared to those in the NC group. In contrast, the mean level of total bilirubin in the HC group was higher than that in the NC group. Clinical predictors of survival Variables which were predictors of overall survival using univariate analysis included T stage (p = 0.034), N stage (p = 0.001), M stage (p = 0.015), TNM stage (p = 0.001), CRP level (p = 0.001), albumin level (p < 0.001), bilirubin level (p = 0.002), platelet count (p = 0.043), lymphocyte percentage (p < 0.001), and treatment methods (p = 0.001) as shown in Table 3. Age, histological type, tumor location, hemoglobin level, and BMI were not significant predictors of survival. The 2-year survival rate for patients with stages I–II disease was 69.5% compared with 22.7% with stages III–IV disease (p = 0.001, Fig. 1a). The HC group had shorter overall survival times compared with the NC group using the univariate analysis. The 2-year survival rate was 78.4% for the NC group compared with 7.8% for the HC group (p < 0.001, Fig. 1b). Fifty-four patients (43.9%) had hypoalbuminemia (albumin <3.5 g/dL). The 2-year survival rates
Table 3 Univariate analysis of prognostic factors. Variables
2-year survival rate (%)
Age (<60/P60 years) Histology (SCC/adenocarcinoma) Tumor location (upper/middle/lower third) T stage (T1-2/T3-4) N stage (N0/N1) M stage (M0/M1) TNM stage (I–II/III–IV) Treatment (trimodality/definitive CCRT/RT only) CRP (<5/P5 mg/L) Albumin (P3.5/<3.5 g/dL) Bilirubin total (61.4/>1.4 mg/dL) Hemoglobin (P10/<10 g/dL) Platelet (P150,000/<150,000/lL) Lymphocyte (P20/<20%) BMI (P18.5/<18.5) GPS (0/1/2)
35.5/35.2 34.9/35.7 16.6/42.0/38.3 67.7/28.0 64.0/22.4 43.6/20.4 69.5/22.7 54.8/26.7/17.0 78.4/7.8 58.5/0.0 42.4/22.2 38.7/0.0 38.8/13.4 51.5/0.0 37.1/35.8 83.6/24.9/0.0
p 0.642 0.532 0.394 0.034 0.001 0.015 0.001 0.001 <0.001 <0.001 0.002 0.425 0.043 <0.001 0.178 <0.001
Univariate survival analysis was performed using the Kaplan–Meier method, and the difference between survival curves was tested by log-rank test. SCC: squamous cell carcinoma; CCRT: concurrent chemoradiotherapy; RT: radiotherapy; CRP: Creactive protein; BMI: Body mass index; GPS: Glasgow Prognostic score.
Fig. 1. Overall survival in patients with (a) stages I–II vs. stages III–IV; (b) normal vs. higher pretreatment CRP levels; and (c) normal vs. lower pretreatment albumin (<3.5 g/dL) levels.
for patient without and with hypoalbuminemia were 58.5% and 0.0%, respectively (p < 0.001, Fig. 1c). Multivariate analysis of prognostic factors of significance showed that only CRP (p < 0.001), albumin (p < 0.001) and treatment methods (p = 0.009) were predictors of overall survival (Table 4, Model A). The hazard ratio was 12.116 for the HC group (95% confidence interval [CI], 3.449–42.567) compared to that for the NC group, and 3.916 for patients with albumin less than
274
CRP as a prognostic factor for esophageal cancer
Table 4 Multivariate analysis of prognostic factors of significance in univariate analysis. Variables
Hazard ratio
95% CI
p
Model A (as categorical variables) Treatment (trimodality/definitive CCRT/RT only) CRP (<5/P5 mg/L) Albumin (P3.5/<3.5 g/dL)
2.407 12.116 3.916
1.251–4.632 3.449–42.567 1.867–8.217
0.009 <0.001 <0.001
1.008–1.019 0.252–0.576
<0.001 <0.001
Model B (as continuous variables) CRP (mg/L) Albumin (g/dL)
1.013 0.381
Multivariate survival analysis was performed using the Cox regression model. CCRT: concurrent chemoradiotherapy; RT: radiotherapy; CRP: C-reactive protein.
Discussion
Fig. 2. ROC curve in patients survived more than 12 months. Sensitivity = 72%, specificity = 56%, PPV = 37% and NPV = 85%.
3.5 g/dL (95% CI, 1.867–8.217) compared to those with albumin greater than 3.5 g/dL. (Table 4, Model A). The ROC curve, as shown in Fig. 2, had an area under curve (AUC) of 0.68 (p = 0.003, 95% confidence interval: 0.58–0.78). When the probability was greater than 0.24, a survival of greater than 1 year was predicted with a sensitivity of 72%, specificity of 56%, PPV of 37% and NPV of 85%. The optimum cutoff value for elevated serum CRP levels (>5 vs. >10 mg/L) was compared using the multivariate survival analysis, and the prognostic significance of a threshold of >5 mg/L (p = 0.001) was better than that using a threshold of >10 mg/L (p = 0.009). When the values of serum biomarkers were entered as continuous variables in the Cox regression model (Table 4, Model B), the hazard ratio of death was significantly increased by 1.3% (95% CI, 0.8–1.9%) for each 1-mg/L increment of the CRP value. As for serum albumin, the hazard ratio of death was significantly decreased by 61.9% (95% CI, 42.4–74.8%) for each 1-g/dL increment. The values of serum CRP and albumin remained the most significant independent prognostic factors. Higher serum CRP concentration and hypoalbuminemia were associated with shorter survival in univariate analysis, and were the independent prognostic factors in multivariate analysis. When these two variables were scored as a combined index (GPS), the total score demonstrated a good stratification value for survival (p < 0.001, Table 3). The GPS was also an independent prognostic factor in the multivariate analysis (p < 0.001). Thirty-three of the 123 patients (26.8%) had a normal GPS of 0, while the majority of patients had abnormal GPS of 1 for 45 patients and GPS of 2 for the remaining 45 patients. The 2-year survival was 83.6% for a GPS of 0, 24.9% for a GPS of 1 and 0.0% for a GPS of 2.
Compared to the Western-based population studies, the histology of the majority of the tumors found in our patients was SCC (94.3%), reflecting the prevailing histological type in Asian men. Although there was no difference in survival between SCC and adenocarcinoma histologies in our study, the treatment method used was significant and this may be secondary to treatment selection bias, since patients with better performance score tended to receive aggressive trimodality treatment rather than radiotherapy only. Higher serum CRP concentration and hypoalbuminemia appear to be independent prognostic factors. In our study, the serum concentration of albumin correlated well with overall survival, while the serum concentration of CRP was inversely related to survival. The association between an elevated CRP concentration and a reduced survival in patients with esophageal cancer is influenced by many confounding factors. First, the presence of a systemic inflammatory response and an accompanying nutritional decline reduces patient tolerance to treatment toxicities and diminishes compliance with treatment [21]. Moreover, the association of elevated serum CRP levels with poor prognosis is attributable to an impaired T-lymphocytic response to cancer cells [22]. In our study, patients in the HC group had more severe lymphocytopenia than those in the NC group suggesting that the host had either a local or a systemic infection. While acute bacterial infections are usually associated with a marked elevation of serum CRP concentration (>10 mg/dL) within 8–12 h, persistent elevation of serum CRP concentration at a moderate degree (<10 mg/dL) has been observed in patients with cancers [23]. It has been suggested that the mild degree of tissue stress due to the presence of cancer cells results in an inflammatory response, and becomes the stimulus for CRP production. The prognosis of patients with esophageal cancer remains poor even with active trimodality treatment, because they are often in a state of impaired immunity and poor nutrition [9]. Up to 83.3% of the patients who had an elevated serum CRP concentration in our study also had hypoalbuminemia. This is consistent with the observation that the development of hypoalbuminemia is often secondary to a systemic inflammatory response [6]. The Glasgow Prognostic score (GPS), which is based on a combination of elevated CRP and hypoalbuminemia levels, is indicative of both an underlying systemic inflammatory response and a nutritional decline. We confirm the importance of GPS in predicting the survival in patients with esophageal cancer treated with radiotherapy. When the relationship between the GPS and 2-year survival rate was examined, there was approximately a three-fold decrease in the survival rate between those patients with a GPS of 0 (83.6%) and those with a GPS of 1 (24.9%). This suggests that there is a subgroup of patients who derive little benefit from a stressful treatment method such as an esophagectomy. Therefore, stratification
C.-Y. Wang et al. / Radiotherapy and Oncology 92 (2009) 270–275
of patients based on this simple and reliable prognostic grouping would aid in the selection of an optimum treatment strategy. Although serum levels of CRP and albumin or GPS have the potential to aid treatment selection between trimodality therapy and definitive CCRT, it should be interpreted cautiously and should not replace the standard TNM staging system. Serial measurements of serum CRP levels to identify patients at risk for recurrence following trimodality treatment or definitive CCRT warrant further investigation. Since our work suggests that serum CRP and albumin are independent prognostic indicators for esophageal cancer, further research should focus on tailoring therapy to specific patient subgroups, such as those with elevated CRP and low albumin, to achieve optimum therapeutic effect. Future randomized trials to evaluate the efficacy of definitive CCRT vs. neoadjuvant CCRT and surgery could include patients with elevated CRP and low albumin to see if surgery can be avoided in those patients who respond poorly. Furthermore, the incorporation of potentially efficacious molecular targeted therapies such as cetuximab provides new validations sets for this subgroup of patients to improve treatment without markedly increasing toxicity. In summary, higher serum CRP concentrations are associated with advanced esophageal cancer. Pretreatment serum levels of CRP and albumin are easily measurable biomarkers of prognostic significance in esophageal cancer. They can be used in combination with the conventional staging to predict survival in patients with esophageal cancer treated with radiotherapy. References [1] Siewert JR, Ott K. Are squamous and adenocarcinomas of the esophagus the same disease? Semin Radiat Oncol 2007;17:38–44. [2] Department of Health, E.Y., Republic of China (Taiwan). Statistics of Causes of Death, 2005. Taipei: Department of Health, Executive Yuan, Republic of China (Taiwan); 2006. [3] Schneider BJ, Urba SG. Preoperative chemoradiation for the treatment of locoregional esophageal cancer: the standard of care? Semin Radiat Oncol 2007;17:45–52. [4] Suntharalingam M. Definitive chemoradiation in the management of locally advanced esophageal cancer. Semin Radiat Oncol 2007;17:22–8. [5] McMillan DC, Canna K, McArdle CS. Systemic inflammatory response predicts survival following curative resection of colorectal cancer. Br J Surg 2003;90:215–9. [6] McMillan DC, Elahi MM, Sattar N, Angerson WJ, Johnstone J, McArdle CS. Measurement of the systemic inflammatory response predicts cancer-specific and non-cancer survival in patients with cancer. Nutr Cancer 2001;41:64–9.
275
[7] Crozier JE, McKee RF, McArdle CS, et al. The presence of a systemic inflammatory response predicts poorer survival in patients receiving adjuvant 5-FU chemotherapy following potentially curative resection for colorectal cancer. Br J Cancer 2006;94:1833–6. [8] Ikeda M, Natsugoe S, Ueno S, Baba M, Aikou T. Significant host- and tumorrelated factors for predicting prognosis in patients with esophageal carcinoma. Ann Surg 2003;238:197–202. [9] Nozoe T, Saeki H, Sugimachi K. Significance of preoperative elevation of serum C-reactive protein as an indicator of prognosis in esophageal carcinoma. Am J Surg 2001;182:197–201. [10] Shimada H, Nabeya Y, Okazumi S, et al. Elevation of preoperative serum Creactive protein level is related to poor prognosis in esophageal squamous cell carcinoma. J Surg Oncol 2003;83:248–52. [11] Crumley AB, McMillan DC, McKernan M, Going JJ, Shearer CJ, Stuart RC. An elevated C-reactive protein concentration, prior to surgery, predicts poor cancer-specific survival in patients undergoing resection for gastrooesophageal cancer. Br J Cancer 2006;94:1568–71. [12] Hashimoto K, Ikeda Y, Korenaga D, et al. The impact of preoperative serum Creactive protein on the prognosis of patients with hepatocellular carcinoma. Cancer 2005;103:1856–64. [13] Kodama J, Miyagi Y, Seki N, et al. Serum C-reactive protein as a prognostic factor in patients with epithelial ovarian cancer. Eur J Obstet Gynecol Reprod Biol 1999;82:107–10. [14] Nakanishi H, Araki N, Kudawara I, et al. Clinical implications of serum Creactive protein levels in malignant fibrous histiocytoma. Int J Cancer 2002;99:167–70. [15] Forrest LM, McMillan DC, McArdle CS, Angerson WJ, Dunlop DJ. Evaluation of cumulative prognostic scores based on the systemic inflammatory response in patients with inoperable non-small-cell lung cancer. Br J Cancer 2003;89:1028–30. [16] Murri AM, Bartlett JM, Canney PA, Doughty JC, Wilson C, McMillan DC. Evaluation of an inflammation-based prognostic score (GPS) in patients with metastatic breast cancer. Br J Cancer 2006;94:227–30. [17] Crumley AB, McMillan DC, McKernan M, McDonald AC, Stuart RC. Evaluation of an inflammation-based prognostic score in patients with inoperable gastrooesophageal cancer. Br J Cancer 2006;94:637–41. [18] Ramsey S, Lamb GW, Aitchison M, Graham J, McMillan DC. Evaluation of an inflammation-based prognostic score in patients with metastatic renal cancer. Cancer 2007;109:205–12. [19] McMillan DC, Crozier JE, Canna K, Angerson WJ, McArdle CS. Evaluation of an inflammation-based prognostic score (GPS) in patients undergoing resection for colon and rectal cancer. Int J Colorectal Dis 2007;22:881–6. [20] Glen P, Jamieson NB, McMillan DC, Carter R, Imrie CW, McKay CJ. Evaluation of an inflammation-based prognostic score in patients with inoperable pancreatic cancer. Pancreatology 2006;6:450–3. [21] Scott HR, McMillan DC, Forrest LM, Brown DJ, McArdle CS, Milroy R. The systemic inflammatory response, weight loss, performance status and survival in patients with inoperable non-small cell lung cancer. Br J Cancer 2002;87:264–7. [22] Schumacher K, Haensch W, Roefzaad C, Schlag PM. Prognostic significance of activated CD8(+) T cell infiltrations within esophageal carcinomas. Cancer Res 2001;61:3932–6. [23] Kushner I, Rzewnicki D, Samols D. What does minor elevation of C-reactive protein signify? Am J Med 2006;119. 166 e117–128.
Contents lists available at ScienceDirect
Radiotherapy and Oncology journal homepage: www.thegreenjournal.com
Radiotherapy of esophagus cancer
Higher serum C-reactive protein concentration and hypoalbuminemia are poor prognostic indicators in patients with esophageal cancer undergoing radiotherapy Chang-Yu Wang a,b,c, Ming-Jang Hsieh d, Yi-Chun Chiu e, Shau-Hsuan Li f,*, Hurng-Wern Huang b, Fu-Min Fang c, Yu-Jie Huang c a
National Kaohsiung University of Applied Sciences, Taiwan Institute of Biomedical Sciences, National Sun Yat-Sen University, Taiwan Department of Radiation Oncology, Chang Gung Memorial Hospital-Kaohsiung Medical Center, Chang Gung University College of Medicine, Kaohsiung, Taiwan d Department of Thoracic & Cardiovascular Surgery, Chang Gung Memorial Hospital-Kaohsiung Medical Center, Chang Gung University College of Medicine, Kaohsiung, Taiwan e Division of Hepato-Gastroenterology, Chang Gung Memorial Hospital-Kaohsiung Medical Center, Chang Gung University College of Medicine, Kaohsiung, Taiwan f Department of Hematology-Oncology, Chang Gung Memorial Hospital-Kaohsiung Medical Center, Chang Gung University College of Medicine, Kaohsiung, Taiwan b c
a r t i c l e
i n f o
Article history: Received 16 February 2008 Received in revised form 12 November 2008 Accepted 5 January 2009 Available online 3 February 2009 Keywords: Esophageal cancer Radiotherapy C-reactive protein (CRP) Albumin Glasgow Prognostic score (GPS)
a b s t r a c t Background and purpose: We evaluate if C-reactive protein (CRP) is an objective biomarker of esophageal cancer in patients undergoing radiotherapy. Materials and methods: Between November 2002 and July 2007, 123 patients undergoing radiotherapy for newly diagnosed esophageal cancer were enrolled. Serum CRP concentration was measured before the initiation of treatment. The relationship between serum CRP levels and other relevant variables such as body mass index, white blood cell count, platelet count, bilirubin, aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase (ALP), cholesterol, hemoglobin, and albumin levels was also analyzed. Results: Eighty-one patients of the 123 patients enrolled (65.9%) had elevated CRP levels (P5 mg/L). The 2-year survival for patients with CRP P5 mg/L was 7.8% compared to 78.4% for patients with CRP <5 mg/ L. Hypoalbuminemia (albumin <3.5 g/dL) was also related to shorter survival using univariate analysis. Multivariate analysis demonstrated that only higher serum CRP concentration and hypoalbuminemia were independent prognostic indicators for survival of patients with esophageal cancer. Conclusions: Pretreatment serum levels of CRP and albumin are objective, easily measurable biomarkers which can be used in combination with conventional staging to accurately predict survival in patients with esophageal cancer treated with radiotherapy. Ó 2009 Elsevier Ireland Ltd. All rights reserved. Radiotherapy and Oncology 92 (2009) 270–275
Esophageal cancer is the sixth most common cause of cancer death among men in Asian countries. Ninety percent of all esophageal cancer in Asian men is squamous cell carcinoma (SCC), which is in contrast to the percentages found in men from western countries [1,2]. Esophageal cancer is usually diagnosed at an advanced stage, making curative surgical resection feasible for only 30–40% of patients [3]. Therefore, neoadjuvant concurrent chemoradiotherapy (CCRT) plays an important role in managing patients with unresectable disease at the time of initial diagnosis. Trimodality treatment provides significant improvement in outcome, which is attributable to the combination of neoadjuvant CCRT with surgery for patients with good performance status. For patients who can-
* Corresponding author. Address: Esophageal Cancer Combined Clinic, Department of Radiation Oncology, Chang Gung Memorial Hospital-Kaohsiung Medical Center, No. 123, Dapi Rd., Niaosong Township, Kaohsiung County 833, Taiwan. E-mail address: [email protected] (S.-H. Li). 0167-8140/$ - see front matter Ó 2009 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.radonc.2009.01.002
not tolerate the surgical procedure, CCRT is provided as an alternative [4]. Since the optimal treatment strategy for advanced esophageal cancer is not well established, there is a need for prognostic markers that can improve patient selection for the various treatment strategies. If surgical resection (trimodality treatment) is not planned, the break in the radiotherapy treatment regimen that is usually required for preoperative surgical evaluation can be avoided, allowing for the completion of CCRT as soon as possible. Systemic inflammation has been associated with cancers, and many patients have an acute phase protein response that can manifest as an elevation in serum C-reactive protein (CRP) concentration [5–7]. There is increasing evidence that an elevated serum CRP level before surgery has prognostic value in patients with resectable esophageal cancer [8–10]. CRP is also a prognostic factor in patients with hepatocellular carcinoma, breast cancer, epithelial ovarian cancer, and malignant fibrous histiocytoma [11–14]. There
C.-Y. Wang et al. / Radiotherapy and Oncology 92 (2009) 270–275
had been much interest in establishing CRP-based prognostic scoring systems for patients with cancer. The Glasgow Prognostic score (GPS), which is based on a combination of elevated CRP and hypoalbuminemia levels, is useful in predicting the survival of patients with non-small-cell lung cancer, breast cancer, colorectal cancer, pancreatic cancer, renal cancer, and gastro-esophageal cancer [15–20]. The purpose of our study is to identify similar serum biomarkers for esophageal cancer. We measure the pretreatment serum levels of CRP as well as the relationship between serum CRP levels and other relevant variables such as body mass index (BMI), white blood cell count, platelet count, bilirubin, aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase (ALP), total cholesterol, hemoglobin, and albumin levels. We then analyze their value in predicting the survival of patients with esophageal cancer who have undergone radiotherapy. Materials and methods Patient population and measurement of CRP Between November 2002 and July 2007, patients with newly diagnosed esophageal SCC or adenocarcinoma undergoing radiotherapy were included in our study. This study was approved by the Institutional Review Board of Chang Gung Memorial Hospital, and informed consent was obtained according to established guidelines. Serum concentration of CRP was measured using the immunonephelometry method on BN II nephelometer (Dade Behring, Marburg, Germany) before the initiation of radiotherapy. Our study used a threshold value of 5 mg/L for serum CRP as suggested by the assay manufacturer. Serum levels less than 5 mg/L were considered normal in this assay. Similar studies from Japan showed the prognostic value of an elevated CRP concentration based on the same threshold for CRP in patients undergoing surgical resection for esophageal SCC [8–10]. However, the studies of gastrointestinal cancer from the UK used a higher CRP threshold (>10 mg/L) than that used in our study [11,17]. Diverse assays in these studies accounts for the use of different CRP cutoffs. The best cutoff values for elevated serum CRP levels (>5 vs. >10 mg/L) were assessed in our study using multivariate survival analysis. In addition, the measurements of serum CRP were repeated after one week if there were initial signs of infection including fever (>38 °C) or WBC count greater than 12,000/mm3. The lowest CRP level was then used for analysis. Pretreatment evaluation Patients were evaluated by a multidisciplinary team including a thoracic surgeon, a medical oncologist, a radiation oncologist, and a gastroenterologist. Patients were staged based on the chest radiograph, barium esophagram, panendoscopy, computed tomography (CT) scan of the chest and abdomen, endoscopic ultrasound (EUS), hepatobiliary ultrasound, and bone scan. Bronchoscopy was performed whenever tracheoesophageal fistula was suspected. Positron emission tomography (PET) was performed if occult distant metastases were suspected. The clinical TNM stage was determined according to the 2002 American Joint Committee on Cancer staging system. Treatment protocol Patients classified as having unresectable lesions were treated with two cycles of cisplatin and 5-fluorouracil (5-FU)-based chemotherapy and radiotherapy, concurrently. Each cycle lasted between 4 and 6 weeks. However, patients with poor performance status did not receive chemotherapy, and were treated with radiotherapy only.
271
Three-dimensional conformal radiotherapy (3DCRT) via a fourfield technique was used in most cases. The gross tumor volume (GTV) was defined based on the CT scan and esophagram, or PET scan, if available. The clinical target volume (CTV) covered GTV with 3 cm superior and inferior margins, and 1 cm radial margins. The planning target volume (PTV) was generated from the expansion of CTV by a 1 cm margin. For upper third and lower third tumor sites, the supraclavicular lymph nodes and celiac lymph nodes were also included for prophylactic irradiation. The prescribed dose was administered to the 95% isodose line, covering PTV for 3600 cGy in 20 fractions. The maximum dose to the spinal cord was limited to 4500 cGy. No more than 40% of the normal lung tissue received a radiation dose of 2000 cGy. Response assessment before esophagectomy Within 1–2 weeks following the end of neoadjuvant CCRT, CT scan and EUS were performed to assess clinical response. The multidisciplinary team reviewed the clinical information to determine if the lesions were resectable. If the lesions were classified as resectable, a trimodality treatment was advised by performing esophagectomy approximately 4–6 weeks after the end of neoadjuvant CCRT. Pathologic complete response (pCR) was defined as the complete disappearance of all viable cancer cells in all surgical specimens including the primary esophageal tumor and lymph nodes. Patients who did not achieve a pathologic complete response received a radiotherapy boost to the tumor bed of an additional 3060 cGy in 17 fractions along with two cycles of cisplatin and 5-FU-based chemotherapy. Patients categorized as unresectable, or who refused esophagectomy after neoadjuvant CCRT, received a further radiotherapy boost up to a total dose of 6660 cGy and another two to four cycles of chemotherapy concurrently to complete the course of CCRT. Statistical analysis Continuous data were presented as mean ± standard deviation (SD). The chi-square test, Student’s t test, and Mann–Whitney U test were used to compare data between the two groups. Survival time was measured from the start of radiotherapy until death or the last date of follow-up. The Kaplan–Meier method was used for univariate survival analysis, and the difference between survival curves was tested by a log-rank test. The Cox regression model was used for the multivariate survival analysis of the clinicopathological variables. Furthermore, the receiver operating characteristic (ROC) curve was employed to determine the positive predictive value (PPV) and negative predictive value (NPV) in patients who survived more than 12 months.
Results Patient characteristics One hundred and twenty males and 3 females, whose ages ranged from 34 to 81 years (median age of 54 years at diagnosis), were enrolled in our study. The clinicopathological characteristics of the patients are shown in Table 1. Histologically, 116 (94.3%) of the patients’ tumors were classified as SCC and the remaining 7 (5.7%) were adenocarcinoma. Tumors were located in the upper third of the esophagus in 19 (15.4%) patients; in the middle third of the esophagus in 65 (52.8%), and in the lower third in 39 (31.7%). According to the TNM classification, six patients were classified as stage I, (4.9%), 16 were classified as stage II (13.0%), 57 were
272
CRP as a prognostic factor for esophageal cancer
Table 1 Relationship between serum CRP levels and clinical characteristics. Mean CRP level mg/L
p
CRP <5 mg/L
CRP P5 mg/L
Patients (n = 42)
Patients (n = 81)
p
Sex Male Female
41.0 ± 55.3 38.7 ± 53.7
0.730
42 (100.0%) 0 (0%)
78 (96.3%) 3 (3.7%)
0.207
Age <60 P60
41.5 ± 50.9 39.9 ± 63.2
0.828
29 (69.0%) 13 (31.0%)
53 (65.4%) 28 (34.6%)
0.687
Histology SCC Adenocarcinoma
38.8 ± 54.4 77.4 ± 56.1
0.025
42 (100.0%) 0 (0.0%)
74 (91.4%) 7 (8.6%)
0.050
Tumor location Upper third Middle third Lower third
67.1 ± 67.1 34.7 ± 51.4 36.2 ± 50.8
0.042
4 (9.5%) 21 (50.0%) 17 (40.5%)
15 (18.5%) 44 (54.3%) 22 (27.2%)
0.213
T stage T1-2 T3-4
38.6 ± 48.5 41.3 ± 56.0
0.572
6 (14.3%) 36 (85.7%)
8 (9.9%) 73 (90.1%)
0.465
N stage N0 N1
30.9 ± 48.7 43.5 ± 56.5
0.034
13 (31.0%) 29 (69.0%)
12 (14.8%) 69 (85.2%)
0.035
M stage M0 M1
34.1 ± 56.6 53.3 ± 50.5
<0.001
37 (88.1%) 5 (11.9%)
42 (51.9%) 39 (48.1%)
<0.001
TNM stage I–II III–IV
27.4 ± 48.2 43.9 ± 56.2
0.005
14 (33.3%) 28 (66.7%)
8 (9.9%) 73 (90.1%)
0.001
Treatment Trimodality Definitive CCRT RT only
20.4 ± 33.1 49.2 ± 57.6 64.5 ± 81.0
0.003
21 (50.0%) 19 (45.2%) 2 (4.8%)
20 (24.7%) 52 (64.2%) 9 (11.1%)
0.016
Continuous data were expressed as mean ± standard deviation. The chi-square test, Student’s t test, and Mann–Whitney U test were used to compare the data between groups. CRP: C-reactive protein; SCC: squamous cell carcinoma; CCRT: concurrent chemoradiotherapy; RT: radiotherapy.
classified as stage III (46.3%), and 44 were as stage IV (35.8%). Those patients with stages I and II disease were considered unsuitable for surgery because of comorbidity. Forty-one (33.3%) of the 123 patients underwent trimodality treatment, 71 (57.7%) patients received definitive CCRT, and the remaining 11 (8.9%) patients were treated with RT only. Of the 41 patients receiving trimodality treatment, 22 (53.7%) also underwent postoperative adjuvant CCRT. The median duration of followup for patients who were alive at the time of the most recent follow-up was 23 months (with a range of 3–57 months). Eight (6.5%) of the 123 patients died of tracheoesophageal fistula with severe pneumonia or massive hemorrhage during the course of neoadjuvant CCRT. Overall, 115 (93.5%) patients completed the planned treatment protocol.
Correlation between CRP levels and clinicopathological parameters Patients were classified as normal CRP (NC) group (n = 42) if their CRP serum level <5 mg/L. If their CRP serum level P5 mg/L, they were classified as the higher CRP (HC) group (n = 81). Patients in the HC group had significantly higher mean CRP levels than those in NC group (60.68 ± 58.92 mg/L vs. 2.98 ± 0.68 mg/L, p < 0.001). Patients with more advanced disease including lymph node metastases (N1), distant metastases (M1), and stages III–IV had significantly higher mean CRP levels than those with early stage cancer (Table 1). There was a good correlation between the treatment method (trimodality vs. CCRT vs. RT only) and the mean CRP level (p = 0.003). In addition, patients with the upper third esophageal cancer tended to have higher mean CRP levels than
Table 2 Relationship between levels of serum CRP and relevant continuous variables.
White blood cell (WBC, count/lL) Hemoglobin (g/dL) Platelet (count/lL) Lymphocyte (%) Albumin (g/dL) Aspartate aminotransferase (AST, U/L) Alanine aminotransferase (ALT, U/L) Alkaline phosphatase (ALP, U/L) Bilirubin total (mg/dL) Total cholesterol (mg/dL) Body mass index (BMI)
CRP <5 mg/L
CRP P5 mg/L
p
7324 ± 2177 13.5 ± 1.7 257690 ± 67942 25.5 ± 8.9 3.8 ± 0.6 26.6 ± 17.5 23.7 ± 40.1 80.3 ± 26.1 0.6 ± 0.3 196.9 ± 46.1 21.0 ± 2.8
8141 ± 2686 12.3 ± 2.1 263988 ± 118904 19.7 ± 10.5 3.2 ± 0.7 31.7 ± 21.3 24.6 ± 20.1 92.7 ± 36.2 0.9 ± 0.7 180.9 ± 47.4 21.0 ± 4.0
0.092 0.001 0.711 0.003 <0.001 0.189 0.870 0.056 0.034 0.140 0.985
Continuous data were presented as mean ± standard deviation. Student’s t test and Mann–Whitney U test were used to compare the data between groups. CRP: C-reactive protein.
C.-Y. Wang et al. / Radiotherapy and Oncology 92 (2009) 270–275
273
those with the middle and the lower third esophageal cancer (p = 0.042). The relationship between serum CRP levels and clinicopathological variables was analyzed (n = 123, Table 1). The proportion of lymph node metastases (N1) in the HC group was significantly higher than the NC group or 85.2% (69 of 81) compared to 69.0% (29 of 42; p = 0.035). Seventy-three of the 101 patients (72.3%) with stages III–IV disease had serum CRP levels P5 mg/L, whereas only 8 of the 22 patients (36.4%) with stages I–II disease had CRP levels P5 mg/L (p = 0.001). Thirty-seven of the 42 patients (88.1%) without distant metastases (M0) had normal CRP levels, compared to only 42 of the 81 patients (51.9%) with distant metastases (M1) (p < 0.001). No significant difference in sex, age, histological type, tumor location, or T stage was found between the HC and the NC group. The relationship between serum CRP levels and relevant continuous variables is shown in Table 2. There was no significant difference in BMI, white blood cell counts, platelet counts, AST, ALT, ALP, and total cholesterol levels between NC and HC groups. Patients in the HC group had lower lymphocyte percentages, hemoglobin, and albumin levels compared to those in the NC group. In contrast, the mean level of total bilirubin in the HC group was higher than that in the NC group. Clinical predictors of survival Variables which were predictors of overall survival using univariate analysis included T stage (p = 0.034), N stage (p = 0.001), M stage (p = 0.015), TNM stage (p = 0.001), CRP level (p = 0.001), albumin level (p < 0.001), bilirubin level (p = 0.002), platelet count (p = 0.043), lymphocyte percentage (p < 0.001), and treatment methods (p = 0.001) as shown in Table 3. Age, histological type, tumor location, hemoglobin level, and BMI were not significant predictors of survival. The 2-year survival rate for patients with stages I–II disease was 69.5% compared with 22.7% with stages III–IV disease (p = 0.001, Fig. 1a). The HC group had shorter overall survival times compared with the NC group using the univariate analysis. The 2-year survival rate was 78.4% for the NC group compared with 7.8% for the HC group (p < 0.001, Fig. 1b). Fifty-four patients (43.9%) had hypoalbuminemia (albumin <3.5 g/dL). The 2-year survival rates
Table 3 Univariate analysis of prognostic factors. Variables
2-year survival rate (%)
Age (<60/P60 years) Histology (SCC/adenocarcinoma) Tumor location (upper/middle/lower third) T stage (T1-2/T3-4) N stage (N0/N1) M stage (M0/M1) TNM stage (I–II/III–IV) Treatment (trimodality/definitive CCRT/RT only) CRP (<5/P5 mg/L) Albumin (P3.5/<3.5 g/dL) Bilirubin total (61.4/>1.4 mg/dL) Hemoglobin (P10/<10 g/dL) Platelet (P150,000/<150,000/lL) Lymphocyte (P20/<20%) BMI (P18.5/<18.5) GPS (0/1/2)
35.5/35.2 34.9/35.7 16.6/42.0/38.3 67.7/28.0 64.0/22.4 43.6/20.4 69.5/22.7 54.8/26.7/17.0 78.4/7.8 58.5/0.0 42.4/22.2 38.7/0.0 38.8/13.4 51.5/0.0 37.1/35.8 83.6/24.9/0.0
p 0.642 0.532 0.394 0.034 0.001 0.015 0.001 0.001 <0.001 <0.001 0.002 0.425 0.043 <0.001 0.178 <0.001
Univariate survival analysis was performed using the Kaplan–Meier method, and the difference between survival curves was tested by log-rank test. SCC: squamous cell carcinoma; CCRT: concurrent chemoradiotherapy; RT: radiotherapy; CRP: Creactive protein; BMI: Body mass index; GPS: Glasgow Prognostic score.
Fig. 1. Overall survival in patients with (a) stages I–II vs. stages III–IV; (b) normal vs. higher pretreatment CRP levels; and (c) normal vs. lower pretreatment albumin (<3.5 g/dL) levels.
for patient without and with hypoalbuminemia were 58.5% and 0.0%, respectively (p < 0.001, Fig. 1c). Multivariate analysis of prognostic factors of significance showed that only CRP (p < 0.001), albumin (p < 0.001) and treatment methods (p = 0.009) were predictors of overall survival (Table 4, Model A). The hazard ratio was 12.116 for the HC group (95% confidence interval [CI], 3.449–42.567) compared to that for the NC group, and 3.916 for patients with albumin less than
274
CRP as a prognostic factor for esophageal cancer
Table 4 Multivariate analysis of prognostic factors of significance in univariate analysis. Variables
Hazard ratio
95% CI
p
Model A (as categorical variables) Treatment (trimodality/definitive CCRT/RT only) CRP (<5/P5 mg/L) Albumin (P3.5/<3.5 g/dL)
2.407 12.116 3.916
1.251–4.632 3.449–42.567 1.867–8.217
0.009 <0.001 <0.001
1.008–1.019 0.252–0.576
<0.001 <0.001
Model B (as continuous variables) CRP (mg/L) Albumin (g/dL)
1.013 0.381
Multivariate survival analysis was performed using the Cox regression model. CCRT: concurrent chemoradiotherapy; RT: radiotherapy; CRP: C-reactive protein.
Discussion
Fig. 2. ROC curve in patients survived more than 12 months. Sensitivity = 72%, specificity = 56%, PPV = 37% and NPV = 85%.
3.5 g/dL (95% CI, 1.867–8.217) compared to those with albumin greater than 3.5 g/dL. (Table 4, Model A). The ROC curve, as shown in Fig. 2, had an area under curve (AUC) of 0.68 (p = 0.003, 95% confidence interval: 0.58–0.78). When the probability was greater than 0.24, a survival of greater than 1 year was predicted with a sensitivity of 72%, specificity of 56%, PPV of 37% and NPV of 85%. The optimum cutoff value for elevated serum CRP levels (>5 vs. >10 mg/L) was compared using the multivariate survival analysis, and the prognostic significance of a threshold of >5 mg/L (p = 0.001) was better than that using a threshold of >10 mg/L (p = 0.009). When the values of serum biomarkers were entered as continuous variables in the Cox regression model (Table 4, Model B), the hazard ratio of death was significantly increased by 1.3% (95% CI, 0.8–1.9%) for each 1-mg/L increment of the CRP value. As for serum albumin, the hazard ratio of death was significantly decreased by 61.9% (95% CI, 42.4–74.8%) for each 1-g/dL increment. The values of serum CRP and albumin remained the most significant independent prognostic factors. Higher serum CRP concentration and hypoalbuminemia were associated with shorter survival in univariate analysis, and were the independent prognostic factors in multivariate analysis. When these two variables were scored as a combined index (GPS), the total score demonstrated a good stratification value for survival (p < 0.001, Table 3). The GPS was also an independent prognostic factor in the multivariate analysis (p < 0.001). Thirty-three of the 123 patients (26.8%) had a normal GPS of 0, while the majority of patients had abnormal GPS of 1 for 45 patients and GPS of 2 for the remaining 45 patients. The 2-year survival was 83.6% for a GPS of 0, 24.9% for a GPS of 1 and 0.0% for a GPS of 2.
Compared to the Western-based population studies, the histology of the majority of the tumors found in our patients was SCC (94.3%), reflecting the prevailing histological type in Asian men. Although there was no difference in survival between SCC and adenocarcinoma histologies in our study, the treatment method used was significant and this may be secondary to treatment selection bias, since patients with better performance score tended to receive aggressive trimodality treatment rather than radiotherapy only. Higher serum CRP concentration and hypoalbuminemia appear to be independent prognostic factors. In our study, the serum concentration of albumin correlated well with overall survival, while the serum concentration of CRP was inversely related to survival. The association between an elevated CRP concentration and a reduced survival in patients with esophageal cancer is influenced by many confounding factors. First, the presence of a systemic inflammatory response and an accompanying nutritional decline reduces patient tolerance to treatment toxicities and diminishes compliance with treatment [21]. Moreover, the association of elevated serum CRP levels with poor prognosis is attributable to an impaired T-lymphocytic response to cancer cells [22]. In our study, patients in the HC group had more severe lymphocytopenia than those in the NC group suggesting that the host had either a local or a systemic infection. While acute bacterial infections are usually associated with a marked elevation of serum CRP concentration (>10 mg/dL) within 8–12 h, persistent elevation of serum CRP concentration at a moderate degree (<10 mg/dL) has been observed in patients with cancers [23]. It has been suggested that the mild degree of tissue stress due to the presence of cancer cells results in an inflammatory response, and becomes the stimulus for CRP production. The prognosis of patients with esophageal cancer remains poor even with active trimodality treatment, because they are often in a state of impaired immunity and poor nutrition [9]. Up to 83.3% of the patients who had an elevated serum CRP concentration in our study also had hypoalbuminemia. This is consistent with the observation that the development of hypoalbuminemia is often secondary to a systemic inflammatory response [6]. The Glasgow Prognostic score (GPS), which is based on a combination of elevated CRP and hypoalbuminemia levels, is indicative of both an underlying systemic inflammatory response and a nutritional decline. We confirm the importance of GPS in predicting the survival in patients with esophageal cancer treated with radiotherapy. When the relationship between the GPS and 2-year survival rate was examined, there was approximately a three-fold decrease in the survival rate between those patients with a GPS of 0 (83.6%) and those with a GPS of 1 (24.9%). This suggests that there is a subgroup of patients who derive little benefit from a stressful treatment method such as an esophagectomy. Therefore, stratification
C.-Y. Wang et al. / Radiotherapy and Oncology 92 (2009) 270–275
of patients based on this simple and reliable prognostic grouping would aid in the selection of an optimum treatment strategy. Although serum levels of CRP and albumin or GPS have the potential to aid treatment selection between trimodality therapy and definitive CCRT, it should be interpreted cautiously and should not replace the standard TNM staging system. Serial measurements of serum CRP levels to identify patients at risk for recurrence following trimodality treatment or definitive CCRT warrant further investigation. Since our work suggests that serum CRP and albumin are independent prognostic indicators for esophageal cancer, further research should focus on tailoring therapy to specific patient subgroups, such as those with elevated CRP and low albumin, to achieve optimum therapeutic effect. Future randomized trials to evaluate the efficacy of definitive CCRT vs. neoadjuvant CCRT and surgery could include patients with elevated CRP and low albumin to see if surgery can be avoided in those patients who respond poorly. Furthermore, the incorporation of potentially efficacious molecular targeted therapies such as cetuximab provides new validations sets for this subgroup of patients to improve treatment without markedly increasing toxicity. In summary, higher serum CRP concentrations are associated with advanced esophageal cancer. Pretreatment serum levels of CRP and albumin are easily measurable biomarkers of prognostic significance in esophageal cancer. They can be used in combination with the conventional staging to predict survival in patients with esophageal cancer treated with radiotherapy. References [1] Siewert JR, Ott K. Are squamous and adenocarcinomas of the esophagus the same disease? Semin Radiat Oncol 2007;17:38–44. [2] Department of Health, E.Y., Republic of China (Taiwan). Statistics of Causes of Death, 2005. Taipei: Department of Health, Executive Yuan, Republic of China (Taiwan); 2006. [3] Schneider BJ, Urba SG. Preoperative chemoradiation for the treatment of locoregional esophageal cancer: the standard of care? Semin Radiat Oncol 2007;17:45–52. [4] Suntharalingam M. Definitive chemoradiation in the management of locally advanced esophageal cancer. Semin Radiat Oncol 2007;17:22–8. [5] McMillan DC, Canna K, McArdle CS. Systemic inflammatory response predicts survival following curative resection of colorectal cancer. Br J Surg 2003;90:215–9. [6] McMillan DC, Elahi MM, Sattar N, Angerson WJ, Johnstone J, McArdle CS. Measurement of the systemic inflammatory response predicts cancer-specific and non-cancer survival in patients with cancer. Nutr Cancer 2001;41:64–9.
275
[7] Crozier JE, McKee RF, McArdle CS, et al. The presence of a systemic inflammatory response predicts poorer survival in patients receiving adjuvant 5-FU chemotherapy following potentially curative resection for colorectal cancer. Br J Cancer 2006;94:1833–6. [8] Ikeda M, Natsugoe S, Ueno S, Baba M, Aikou T. Significant host- and tumorrelated factors for predicting prognosis in patients with esophageal carcinoma. Ann Surg 2003;238:197–202. [9] Nozoe T, Saeki H, Sugimachi K. Significance of preoperative elevation of serum C-reactive protein as an indicator of prognosis in esophageal carcinoma. Am J Surg 2001;182:197–201. [10] Shimada H, Nabeya Y, Okazumi S, et al. Elevation of preoperative serum Creactive protein level is related to poor prognosis in esophageal squamous cell carcinoma. J Surg Oncol 2003;83:248–52. [11] Crumley AB, McMillan DC, McKernan M, Going JJ, Shearer CJ, Stuart RC. An elevated C-reactive protein concentration, prior to surgery, predicts poor cancer-specific survival in patients undergoing resection for gastrooesophageal cancer. Br J Cancer 2006;94:1568–71. [12] Hashimoto K, Ikeda Y, Korenaga D, et al. The impact of preoperative serum Creactive protein on the prognosis of patients with hepatocellular carcinoma. Cancer 2005;103:1856–64. [13] Kodama J, Miyagi Y, Seki N, et al. Serum C-reactive protein as a prognostic factor in patients with epithelial ovarian cancer. Eur J Obstet Gynecol Reprod Biol 1999;82:107–10. [14] Nakanishi H, Araki N, Kudawara I, et al. Clinical implications of serum Creactive protein levels in malignant fibrous histiocytoma. Int J Cancer 2002;99:167–70. [15] Forrest LM, McMillan DC, McArdle CS, Angerson WJ, Dunlop DJ. Evaluation of cumulative prognostic scores based on the systemic inflammatory response in patients with inoperable non-small-cell lung cancer. Br J Cancer 2003;89:1028–30. [16] Murri AM, Bartlett JM, Canney PA, Doughty JC, Wilson C, McMillan DC. Evaluation of an inflammation-based prognostic score (GPS) in patients with metastatic breast cancer. Br J Cancer 2006;94:227–30. [17] Crumley AB, McMillan DC, McKernan M, McDonald AC, Stuart RC. Evaluation of an inflammation-based prognostic score in patients with inoperable gastrooesophageal cancer. Br J Cancer 2006;94:637–41. [18] Ramsey S, Lamb GW, Aitchison M, Graham J, McMillan DC. Evaluation of an inflammation-based prognostic score in patients with metastatic renal cancer. Cancer 2007;109:205–12. [19] McMillan DC, Crozier JE, Canna K, Angerson WJ, McArdle CS. Evaluation of an inflammation-based prognostic score (GPS) in patients undergoing resection for colon and rectal cancer. Int J Colorectal Dis 2007;22:881–6. [20] Glen P, Jamieson NB, McMillan DC, Carter R, Imrie CW, McKay CJ. Evaluation of an inflammation-based prognostic score in patients with inoperable pancreatic cancer. Pancreatology 2006;6:450–3. [21] Scott HR, McMillan DC, Forrest LM, Brown DJ, McArdle CS, Milroy R. The systemic inflammatory response, weight loss, performance status and survival in patients with inoperable non-small cell lung cancer. Br J Cancer 2002;87:264–7. [22] Schumacher K, Haensch W, Roefzaad C, Schlag PM. Prognostic significance of activated CD8(+) T cell infiltrations within esophageal carcinomas. Cancer Res 2001;61:3932–6. [23] Kushner I, Rzewnicki D, Samols D. What does minor elevation of C-reactive protein signify? Am J Med 2006;119. 166 e117–128.