H&E images illustrating the effects of chemoradiation on esophageal adeno- and squamous cell carcinomas. A] TRG I, adenocarcinoma with no or little therapy effect. B] TRG II, residual squamous cell carcinoma and therapy effects (squamous remnant, arrow). C] TRG III, few residual tumor cells (arrows) surrounded by fibrosis of the adventitia and the muscular layer. D] TRG IV, an ulcer is seen directly above the muscular layer showing fibrosis, but no vital tumor cells. E, F] TRG IV, showing mucin lakes and squamous remnants (arrows), respectively, but no vital tumor cells. G, H] lymph nodes displaying mucin and keratin squames, respectively, in the subcapsular sinus (arrow), no tumor cells. I,J] well differentiated adenocarcinoma in the preoperative biopsy (I) , and moderately (to poorly) differentiated adenocarcinoma in the resection specimen (J) , illustrating downgrading after chemoradiation. K, L] likewise, moderately differentiated (keratinizing, arrow) squamous cell carcinoma in the preoperative biopsy (K) , and poorly differentiated squamous cell carcinoma after chemoradiation (L) . A 4x objective was used in A-F, 10x in G-H, and 20x in I-L. 

Contexts in source publication

Context 1

... One patient was hospitalized because of rectal bleeding. Colonoscopy revealed a non-malignant polyp, which was removed. One patient was hospitalized because of vomiting and fever with grade 2 neutropenia. Evaluation with upper GI endoscopy and CT was done after a mean of ten days following the last radiotherapy session. Response evaluation with endoscopy showed a complete response in ten patients (18.9%), a major response in 23 patients (43.4%), a minor response in 11 patients (20.8%), and no response in one patient (1.9%). In five patients response evaluation by endoscopy was not possible (in one patient there was “no pass”, in four patients the baseline endoscopy was performed in another hospital) and in three patients no endoscopy was performed after completing the chemoradiation. Response evaluated with CT showed no complete response or disease progression. In three patients (5.6 % ) a partial response was observed. One patient refused surgery after having completed the chemoradiotherapy. Endoscopy in this patient revealed a complete response. After 12 months of follow-up, a local recurrence of the oesophageal tumour was diagnosed, further workup also revealed supraclavicular lymph nodes. He refused further treatment. Thirteen months later he died from progressive disease. A transhiatal oesophagectomy was performed in 46 of the 52 patients, and a transthoracic resection in 6 patients. The median time between the completion of chemoradiotherapy and surgery was 42 days (range 20 to 74 days). The in hospital postoperative mortality rate was 7.7% (CI 0-15%). Two patients died from systemic complications due to anastomotic leakage, one patient from a cerebral vascular accident one day after surgery, and one from sepsis. Autopsy in the latter patient revealed a prostatitis as the probable focus of the sepsis. Postoperative complications were seen in 38 patients (73%). These complications were mainly pulmonary (42%) or cardiac (13%) (Table 4). Besides the two lethal anastomotic leaks, in 10 of the 48 (20.8%) patients surviving postoperatively an anastomotic leak was seen. In five of them (10.4%) the leakage was a radiological finding on routinely performed contrast swallow postoperatively. In all patients the clinical signs of the leak could be treated conservatively, but in two patients (4.2%) the leak resulted in long-term nutritional support. Twenty-two patients (45.8%) developed an anastomotic stricture requiring endoscopic dilatation (range 1-27, median 7). Eventually, all patients were able to eat solid food. In 13 patients no residual tumour in the resected oesophagus or regional lymph nodes was found, corresponding to a pathological complete response (pCR) rate of 25%. The pathological stages of the other resection specimens were: pT1N0-1M0 in 12 patients (23.1%), pT2N0-1M0 in 6 patients (11.5%), pT3N0-1M0 in 16 patients (30.8%), pT0-3N0- 1M1A in 4 patients (7.7%), and pT1N1M1B in 1 patient (1.9%). In 19 patients (36.5%) a regression grade III, in 14 patients (26.9%) a regression grade II, and in 6 patients (11.5%) a histopathological regression grade I was seen. In 7 of the 18 patients (38.9%) with a pathological T3-stage only scattered tumour cells were found in the resection specimen. A radical resection with no evidence of tumour cells at the resection margins (R0-resection) was obtained in all patients. The lymph node dissection status showed a median of eight nodes (range 0-30), derived from both regional and distant sites. In 13 patients (25%) one or more positive lymph nodes were found (median 2, range 1-6). The N-stage improved from N1, as assessed by EUS, to N0 postoperatively in 19 patients (36.5%). In 4 patients (7.7%) the N0-stage, as assessed by EUS, was changed towards a N1-stage postoperatively. The post-treatment pulmonary function tests (measured 6 months and 1 year after surgery) deteriorated significantly compared to the pre-treatment tests. The total lung capacity (TLC) decreased from 103 percent of the predicted value to 92 percent (p=0.002). The vital capacity (VC) declined from 105 percent of the predicted value to 96 percent (p<0.001). The forced expiratory volume in 1 second (FEV1) decreased from 94 percent of the predicted value to 87 percent (p<0.0001). This decline in pulmonary function tests did not lead to major clinical symptoms. All 54 patients were included in the survival analysis. At the time of evaluation (May 31, 2005) the median follow-up time for all patients was 23.5 months (range 0 to 52 months). The median follow-up time for surviving patients was 31 months (range 11- 52 months). Nineteen of the 54 patients (35.2%) died: 13 due to recurrent cancer, five during treatment (four postoperatively and one sudden death) and one due to a rup- tured aortic aneurysm. The median survival time, however, has not yet been reached. The estimated 1-, 2-, and 3-year survival rates were 82%, 65%, and 56%, respectively. The Kaplan-Meier curve for overall survival is shown in Figure 1. The survival of patients with a pCR was not better than the survival of patients with no pCR. Recurrent disease after surgery was found in 15 patients surviving postoperatively (15/48, 31.2%). Three of them were still alive at the time of analysis. Recurrence was locoregional in seven patients. Distant metastases were found in 14 patients. The patient who refused surgery died from recurrent disease as well (see before). The Kaplan-Meier curve for disease free survival of the patients surviving postoperatively is shown in Figure 1. The patient who died without recurrence was censored at the time of death. Preoperative chemoradiotherapy is nowadays widely used in the treatment of patients with potentially resectable oesophageal cancer. The concept that preoperative chemoradiotherapy may lead to a better tumour control and therefore to a better overall survival is appealing, as 29-43% incomplete resections are performed when patients are treated with surgery alone or with chemotherapy followed by surgery [3-5]. Many studies have reported that after chemoradiotherapy in 10-28% of the patients no tumour cells are found in the resection specimen [9, 22-25]. However, surprisingly few phase III studies have been reported in which preoperative chemoradiotherapy followed by surgery was compared with surgery alone. Meta-analyses of these trials showed a small, if any, effect on survival [6-8]. In addition, the results of a recently reported study were disappointing, showing no survival benefit for those patients treated with preoperative chemoradiotherapy [26]. In most studies, the combination of 5-FU and cisplatin with radiotherapy has been applied. In our study we used paclitaxel and carboplatin with concurrent radiotherapy. Our study showed that preoperative chemoradiotherapy with weekly paclitaxel and carboplatin was well tolerated. All patients completed the chemoradiotherapy as scheduled, without treatment delay or dose reduction. The major non-haematological toxicity was a grade 3 or 4 oesophagitis in 7.5% of the patients. Compared to other studies with chemoradiotherapy in oesophageal cancer, this incidence of grade 3 and 4 oesophagitis is low [25, 27]. The postoperative mortality of this study (7.7%; 95% CI 0-15%) was somewhat higher than the approximately 4% mortality rate found in other trials performed at our institution [4, 28, 29], however, the observed mortality rate still lies within the 95% confidence limits. Postoperative morbidity consisted mainly of pulmonary complications. This high pulmonary complication rate is partly due to the fact that we also scored minor pulmonary complications, such as upper airway infections. Whether preoperative chemoradiotherapy is responsible for a higher pulmonary complication rate cannot be excluded. In a retrospective study of Avendano et al preoperative chemoradiotherapy was associated with an increase risk of pulmonary complications (i.e., duration of me- chanical ventilation) [30]. Whether the decline in pulmonary function tests (TLC, VC, and FEV1) that we observed in our study was due to the chemoradiotherapy is uncertain, as it has also been reported that the TLC and the VC were significantly reduced after an oesophagectomy without preoperative treatment [31]. During follow-up 22 patients required dilatations because of an anastomotic stricture. The dilatations resolved the dysphagia in all patients, eventually all patients had an adequate food intake. Neoadjuvant chemoradiotherapy has not been reported to be associated with a higher stricture formation rate [32]. In this study, the overall and disease free survivals compare favourably with those in other trials of preoperative chemoradiation for oesophageal cancer. With a median follow-up of 23.5 months, the median survival time has not yet been reached. However, such findings should be interpreted with caution, because phase II studies always carry the risk of selection bias. A complete (R0) resection was accomplished in all patients, using a 1 mm cut-off point for circumferential resection margin and this also compares favourably with other studies. The pathologically complete response rate of 25% in our study is consistent with that in other studies using preoperative chemoradiation. A major histomorphological regression was seen in another19 resected specimens. Thus, in a total of 32 patients (61.5%) a major or complete pathological response to preoperative chemoradiation was found. Several studies have shown that a pCR and an R0-resection are associated with a better prognosis [33, 34]. Surprisingly, we were not able to demonstrate a significant survival difference between patients who had a pCR and those who did not have a pCR. Since all operated patients had an R0 resection, a possible adverse effect of an incomplete resection on survival could not be assessed. The 100% complete resection rate and high number of patients with a major or complete pathological response might explain the lack ...

Context 2

... patients response evaluation by endoscopy was not possible (in one patient there was “no pass”, in four patients the baseline endoscopy was performed in another hospital) and in three patients no endoscopy was performed after completing the chemoradiation. Response evaluated with CT showed no complete response or disease progression. In three patients (5.6 % ) a partial response was observed. One patient refused surgery after having completed the chemoradiotherapy. Endoscopy in this patient revealed a complete response. After 12 months of follow-up, a local recurrence of the oesophageal tumour was diagnosed, further workup also revealed supraclavicular lymph nodes. He refused further treatment. Thirteen months later he died from progressive disease. A transhiatal oesophagectomy was performed in 46 of the 52 patients, and a transthoracic resection in 6 patients. The median time between the completion of chemoradiotherapy and surgery was 42 days (range 20 to 74 days). The in hospital postoperative mortality rate was 7.7% (CI 0-15%). Two patients died from systemic complications due to anastomotic leakage, one patient from a cerebral vascular accident one day after surgery, and one from sepsis. Autopsy in the latter patient revealed a prostatitis as the probable focus of the sepsis. Postoperative complications were seen in 38 patients (73%). These complications were mainly pulmonary (42%) or cardiac (13%) (Table 4). Besides the two lethal anastomotic leaks, in 10 of the 48 (20.8%) patients surviving postoperatively an anastomotic leak was seen. In five of them (10.4%) the leakage was a radiological finding on routinely performed contrast swallow postoperatively. In all patients the clinical signs of the leak could be treated conservatively, but in two patients (4.2%) the leak resulted in long-term nutritional support. Twenty-two patients (45.8%) developed an anastomotic stricture requiring endoscopic dilatation (range 1-27, median 7). Eventually, all patients were able to eat solid food. In 13 patients no residual tumour in the resected oesophagus or regional lymph nodes was found, corresponding to a pathological complete response (pCR) rate of 25%. The pathological stages of the other resection specimens were: pT1N0-1M0 in 12 patients (23.1%), pT2N0-1M0 in 6 patients (11.5%), pT3N0-1M0 in 16 patients (30.8%), pT0-3N0- 1M1A in 4 patients (7.7%), and pT1N1M1B in 1 patient (1.9%). In 19 patients (36.5%) a regression grade III, in 14 patients (26.9%) a regression grade II, and in 6 patients (11.5%) a histopathological regression grade I was seen. In 7 of the 18 patients (38.9%) with a pathological T3-stage only scattered tumour cells were found in the resection specimen. A radical resection with no evidence of tumour cells at the resection margins (R0-resection) was obtained in all patients. The lymph node dissection status showed a median of eight nodes (range 0-30), derived from both regional and distant sites. In 13 patients (25%) one or more positive lymph nodes were found (median 2, range 1-6). The N-stage improved from N1, as assessed by EUS, to N0 postoperatively in 19 patients (36.5%). In 4 patients (7.7%) the N0-stage, as assessed by EUS, was changed towards a N1-stage postoperatively. The post-treatment pulmonary function tests (measured 6 months and 1 year after surgery) deteriorated significantly compared to the pre-treatment tests. The total lung capacity (TLC) decreased from 103 percent of the predicted value to 92 percent (p=0.002). The vital capacity (VC) declined from 105 percent of the predicted value to 96 percent (p<0.001). The forced expiratory volume in 1 second (FEV1) decreased from 94 percent of the predicted value to 87 percent (p<0.0001). This decline in pulmonary function tests did not lead to major clinical symptoms. All 54 patients were included in the survival analysis. At the time of evaluation (May 31, 2005) the median follow-up time for all patients was 23.5 months (range 0 to 52 months). The median follow-up time for surviving patients was 31 months (range 11- 52 months). Nineteen of the 54 patients (35.2%) died: 13 due to recurrent cancer, five during treatment (four postoperatively and one sudden death) and one due to a rup- tured aortic aneurysm. The median survival time, however, has not yet been reached. The estimated 1-, 2-, and 3-year survival rates were 82%, 65%, and 56%, respectively. The Kaplan-Meier curve for overall survival is shown in Figure 1. The survival of patients with a pCR was not better than the survival of patients with no pCR. Recurrent disease after surgery was found in 15 patients surviving postoperatively (15/48, 31.2%). Three of them were still alive at the time of analysis. Recurrence was locoregional in seven patients. Distant metastases were found in 14 patients. The patient who refused surgery died from recurrent disease as well (see before). The Kaplan-Meier curve for disease free survival of the patients surviving postoperatively is shown in Figure 1. The patient who died without recurrence was censored at the time of death. Preoperative chemoradiotherapy is nowadays widely used in the treatment of patients with potentially resectable oesophageal cancer. The concept that preoperative chemoradiotherapy may lead to a better tumour control and therefore to a better overall survival is appealing, as 29-43% incomplete resections are performed when patients are treated with surgery alone or with chemotherapy followed by surgery [3-5]. Many studies have reported that after chemoradiotherapy in 10-28% of the patients no tumour cells are found in the resection specimen [9, 22-25]. However, surprisingly few phase III studies have been reported in which preoperative chemoradiotherapy followed by surgery was compared with surgery alone. Meta-analyses of these trials showed a small, if any, effect on survival [6-8]. In addition, the results of a recently reported study were disappointing, showing no survival benefit for those patients treated with preoperative chemoradiotherapy [26]. In most studies, the combination of 5-FU and cisplatin with radiotherapy has been applied. In our study we used paclitaxel and carboplatin with concurrent radiotherapy. Our study showed that preoperative chemoradiotherapy with weekly paclitaxel and carboplatin was well tolerated. All patients completed the chemoradiotherapy as scheduled, without treatment delay or dose reduction. The major non-haematological toxicity was a grade 3 or 4 oesophagitis in 7.5% of the patients. Compared to other studies with chemoradiotherapy in oesophageal cancer, this incidence of grade 3 and 4 oesophagitis is low [25, 27]. The postoperative mortality of this study (7.7%; 95% CI 0-15%) was somewhat higher than the approximately 4% mortality rate found in other trials performed at our institution [4, 28, 29], however, the observed mortality rate still lies within the 95% confidence limits. Postoperative morbidity consisted mainly of pulmonary complications. This high pulmonary complication rate is partly due to the fact that we also scored minor pulmonary complications, such as upper airway infections. Whether preoperative chemoradiotherapy is responsible for a higher pulmonary complication rate cannot be excluded. In a retrospective study of Avendano et al preoperative chemoradiotherapy was associated with an increase risk of pulmonary complications (i.e., duration of me- chanical ventilation) [30]. Whether the decline in pulmonary function tests (TLC, VC, and FEV1) that we observed in our study was due to the chemoradiotherapy is uncertain, as it has also been reported that the TLC and the VC were significantly reduced after an oesophagectomy without preoperative treatment [31]. During follow-up 22 patients required dilatations because of an anastomotic stricture. The dilatations resolved the dysphagia in all patients, eventually all patients had an adequate food intake. Neoadjuvant chemoradiotherapy has not been reported to be associated with a higher stricture formation rate [32]. In this study, the overall and disease free survivals compare favourably with those in other trials of preoperative chemoradiation for oesophageal cancer. With a median follow-up of 23.5 months, the median survival time has not yet been reached. However, such findings should be interpreted with caution, because phase II studies always carry the risk of selection bias. A complete (R0) resection was accomplished in all patients, using a 1 mm cut-off point for circumferential resection margin and this also compares favourably with other studies. The pathologically complete response rate of 25% in our study is consistent with that in other studies using preoperative chemoradiation. A major histomorphological regression was seen in another19 resected specimens. Thus, in a total of 32 patients (61.5%) a major or complete pathological response to preoperative chemoradiation was found. Several studies have shown that a pCR and an R0-resection are associated with a better prognosis [33, 34]. Surprisingly, we were not able to demonstrate a significant survival difference between patients who had a pCR and those who did not have a pCR. Since all operated patients had an R0 resection, a possible adverse effect of an incomplete resection on survival could not be assessed. The 100% complete resection rate and high number of patients with a major or complete pathological response might explain the lack of survival benefit in patients who had a pCR. In conclusion, this study shows that preoperative treatment with weekly paclitaxel and carboplatin with concurrent radiotherapy is well tolerated, with leucopenia and oesophagitis being the most common side effects. After chemoradiotherapy a high rate of radical resections could be achieved and the overall survival looks promising. A randomized phase III trial with this regimen followed by surgery versus surgery alone is now ongoing, which has, up to now, included more than 100 ...

Context 3

... as well as stromal changes such as fibrosis (with or without inflammatory infiltrate), including giant cell granuloma formulation around ghost cells and mucin pools or keratin pearls at the site of the previous tumor [1-4]. On the basis of these changes, the grade of tumor regression can be defined. Regression of the primary tumor can range from the absence of regressive changes to a total response with no vital residual tumor cells. Several morphologic criteria have been defined for various types of cancer to objectively evaluate pathological response to neoadjuvant chemoradiation [1, 2, 5]. The prognosis of resectable esophageal cancer is mainly determined by the achieve- ment of a complete tumor resection (R0 resection), the depth of tumor infiltration, and the presence of lymph node metastasis [6-9]. Generally, the overall survival after neoadjuvant chemoradiotherapy is better in patients with a complete pathological response compared to patients with residual tumor cells [6, 8, 10-12]. Schneider et al. and Brücher et al. found that patients with only a few residual tumor cells within the resection specimen (<10%) had a significant better survival compared with patients with >10% residual tumor cells. In this article we review our experience with neoadjuvant carboplatin and paclitaxel concurrent with radiotherapy in patients with resectable esophageal cancer. We determined the residual tumor cells after preoperative treatment, and correlated the effect of specific pathologic and clinical findings to overall survival, in order to obtain prognostic information. We examined the esophageal biopsies and surgical specimens of all patients who were treated with neoadjuvant chemoradiation followed by surgery for esophageal cancer at the Erasmus Medical Centre between February 2001 and August 2006. Patients with T1N1 or T2-3N0-1 esophageal cancer were treated at our institution with five weekly courses of paclitaxel and carboplatin with concurrent radiotherapy followed by surgery in a phase II or phase III trial as described elsewhere [13]. The clinical data collected included patient age, sex, and preoperative clinical stage (obtained from CT scan and EUS). All haematoxylin and eosin (H&E) slides from preoperative tumor biopsies and from the resection specimens were evaluated by a specialized pathologist (H.v.D.). The preoperative tumor biopsies were viewed for histological tumor type, and grade. The resection specimen was evaluated using a standard protocol, providing information on margins, tumor type, tumor differentiation grade, extension of the tumor, and lymph nodes. If no macroscopically identifiable tumor was present, lesions such as an ulcer or an irregular area covered by mucosa were embedded in total together with surrounding areas in order to adequately judge the presence of residual tumor and therapy effects. The slides from the resection specimens were viewed for the presence of vital tumor cells near the oral, ab-oral and circumferential resection margins, tumor type, tumor differentiation grade, extension of the tumor in the esophageal wall, and number of involved lymph nodes. R0 was defined as histologically tumor-free resection margins with ≥ 1 mm distance between tumor and resection margins. The 6th edition of the International Union Against Cancer (UICC) was used for TNM-classification, tumor grade, and stage grouping [14]. The grading of the therapy response was performed according to Junker et al. as described before [2, 13]. The degree of histomorphological regression in the esophageal wall, i.e. the effect of chemoradiation, was classified into four categories: tumor regression grade (TRG) I: more than 50% vital residual tumor cells; TRG II: 10-50% vital residual tumor; TRG III: less than 10% vital residual tumor cells; TRG IV: complete tumor regression, no evidence of vital tumor cells (Figure 1). Special attention was paid to therapy effects, such as mucin pools and squamous remnants in resection margins and lymph nodes (Figure 1). Patient characteristics are described using tables for categorical data, medians and ranges for continuous variables. Association between clinical and pathological param- eters was evaluated using the chi-square test. Overall survival time was calculated from the time of start of the chemoradiotherapy to the date of death or last follow-up. November 15 th 2007 was the censoring date for survival. Survival curves were estimated using the Kaplan-Meier method, and the log-rank test was used to evaluate the statistical significance of differences. The SPSS statistical package (version 12.0; SPSS Inc., Chicago, IL) was used for all statistical analyses. The 67 patients who were assessed in the current study had a median age of 59 years and the majority was male. Baseline patient characteristics are summarized in Table 1. The morphological changes in the tumor and in non-neoplastic tissue: such as mucin pools, squamous remnants, and stromal changes are shown in Figure 1. In Table 2 pathological characteristics of the resection specimens are summarized. A complete microscopic and macroscopic tumor resection was achieved in 66 of 67 patients (98.5%). In these 66 cases there were no mucin pools or areas containing keratin pearls or squamous remnants in the resection margins. When there were still vital tumor cells left in the resection specimen, the tumor differentiation grade was poorer compared to the tumor differentiation grade in the esophageal biopsies before chemoradiotherapy: Seven patients had differentiation grade 3/4 before chemoradiotherapy versus 27 patients after chemoradiation (p<0.001). TRG IV was achieved in 16 patients (24%), in 10 resection specimens (15%), little or no regressive changes were seen (TRG I). The response to chemoradiotherapy (as defined by TRG III and IV) was significantly higher in patients with squamous cell carcinoma (SCC) than in patients with adenocarcinoma (AC) (68% versus 54%, p=0.001). Poorly differentiated tumors (grade 3/4, esophageal biopsies) achieved significantly more tumor regression grade IV than good or moderately differentiated tumors (53% vs. 15%, p<0.01). Fifty-seven patients (85%) had tumor infiltration into the adventitia (T3) at baseline assessed by EUS, versus 25 patients (37%) after therapy as assessed by pathological analysis of the resection specimen, indicating significant downstaging by chemoradio- therapy (p<0.0001). Likewise, 38 patients (57%) had pre-treatment nodal involvement, as assessed by EUS, versus 21 patients (31%) post-therapy, as assessed by pathological examination of the resected specimen (p=0.006). In 11 of the 29 resected specimens with no pre-treatment nodal involvement (38%) regression effects were identified in the lymph nodes, suggesting downstaging from clinical stage N1 to pathological stage N0. The tumor regression grade of the primary tumor was significantly related to the risk of residual tumor cells in the esophageal wall: pT3 was found in 80% for TRG I, 65% for TRG II, and 25% for TRG III (p=0.01). This also applied for the resected lymph nodes: positive lymph nodes were found in 70% for TRG I, 50% for TRG II, 20% for TRG III, and 14% for TRG IV (p=0.008). After a median follow-up of 35 months (range 2-75), 20 patients relapsed (1 patient had a local/regional relapse, 9 patients developed distant metastases, and 10 patient developed both). The presence of pre-treatment nodal involvement was significantly associated with the development of distant metastases (14% vs. 42%, p=0.02). The median overall survival as calculated by the Kaplan Meier method has not yet been reached (Figure 2A). There was a trend towards a better overall survival in patients with a complete regression, i.e. TRG IV (p=0.09). The median overall survival for patients with TRG I- III was 57 months; the median survival for patients with TRG IV has not yet been reached (Figure 2B). Survival was significantly better in patients who had no nodal involvement before the start of treatment (median survival has not yet been reached) compared to patients who had nodal involvement as assessed by EUS (median survival 39 months, p=0.03) (Figure 3). No difference in overall survival could be found between patients with pathological nodal involvement compared to patients with no pathological nodal involvement, including 11 patients (16.7%) with therapy effects within the lymph nodes. There was no survival difference between patients with AC or SCC. In this study we described the histopathological changes seen after neoadjuvant chemoradiotherapy with paclitaxel and carboplatin in 67 cases of esophageal carcinoma. When vital tumor cells were left in the esophageal wall after chemoradiotherapy, the tumor was graded more frequently as poorly differentiated than before treatment. There are 3 explanations for the worsened differentiation grade in the resected specimens. First, poorly differentiated tumor cells may be more resistant to the cytotoxic insult caused by neoadjuvant therapy. However, TRG IV was more frequently found in pretreatment poorly differentiated tumors, so this does not seem to be a likely explanation. Second, it might be caused by “sampling error” of the pretreatment esophageal biopsies, because the tumor differentiation grade of the resected specimen is determined by the worst differentiation found. Lastly, cytotoxic injury of chemoradiotherapy may induce dedif- ferentiation of tumor cells. Changes in differentiation of tumor cells after chemoradiotherapy have been described before [10, 15, 16] We found complete tumor regression (TRG IV) in 24% of the patients, which corresponds with the rates found in other studies with neoadjuvant chemoradiotherapy in patients with esophageal cancer [17-20]. It has been well known that patients who achieve a pathological response after neoadjuvant chemoradiotherapy have a significantly prolonged survival compared to patients with residual vital tumor cells in the ...

Context 4

... than adenocarcinoma. This was not associated with a survival benefit. Neoadjuvant chemoradiotherapy led to a significant down-staging. Patients with pre-treatment nodal involvement had a significantly worse survival compared to patients with no pretreatment nodal involvement. This applied not for post-treatment nodal involvement. Conclusions. Poorly differentiated esophageal tumors significantly more frequent reach complete tumor regression after neoadjuvant chemoradiotherapy. The presence of nodal involvement as assessed by endoscopic ultrasound is significantly associated with a worse survival. Neoadjuvant chemoradiotherapy for patients with resectable esophageal cancer is increasingly used in an attempt to improve these patients. After chemoradiotherapy several histological changes in the resected specimen can be identified. These changes include cytological alterations such as cytoplasmic vacuolation and/or eosinophilia, nuclear pleomorphism, and necrosis, as well as stromal changes such as fibrosis (with or without inflammatory infiltrate), including giant cell granuloma formulation around ghost cells and mucin pools or keratin pearls at the site of the previous tumor [1-4]. On the basis of these changes, the grade of tumor regression can be defined. Regression of the primary tumor can range from the absence of regressive changes to a total response with no vital residual tumor cells. Several morphologic criteria have been defined for various types of cancer to objectively evaluate pathological response to neoadjuvant chemoradiation [1, 2, 5]. The prognosis of resectable esophageal cancer is mainly determined by the achieve- ment of a complete tumor resection (R0 resection), the depth of tumor infiltration, and the presence of lymph node metastasis [6-9]. Generally, the overall survival after neoadjuvant chemoradiotherapy is better in patients with a complete pathological response compared to patients with residual tumor cells [6, 8, 10-12]. Schneider et al. and Brücher et al. found that patients with only a few residual tumor cells within the resection specimen (<10%) had a significant better survival compared with patients with >10% residual tumor cells. In this article we review our experience with neoadjuvant carboplatin and paclitaxel concurrent with radiotherapy in patients with resectable esophageal cancer. We determined the residual tumor cells after preoperative treatment, and correlated the effect of specific pathologic and clinical findings to overall survival, in order to obtain prognostic information. We examined the esophageal biopsies and surgical specimens of all patients who were treated with neoadjuvant chemoradiation followed by surgery for esophageal cancer at the Erasmus Medical Centre between February 2001 and August 2006. Patients with T1N1 or T2-3N0-1 esophageal cancer were treated at our institution with five weekly courses of paclitaxel and carboplatin with concurrent radiotherapy followed by surgery in a phase II or phase III trial as described elsewhere [13]. The clinical data collected included patient age, sex, and preoperative clinical stage (obtained from CT scan and EUS). All haematoxylin and eosin (H&E) slides from preoperative tumor biopsies and from the resection specimens were evaluated by a specialized pathologist (H.v.D.). The preoperative tumor biopsies were viewed for histological tumor type, and grade. The resection specimen was evaluated using a standard protocol, providing information on margins, tumor type, tumor differentiation grade, extension of the tumor, and lymph nodes. If no macroscopically identifiable tumor was present, lesions such as an ulcer or an irregular area covered by mucosa were embedded in total together with surrounding areas in order to adequately judge the presence of residual tumor and therapy effects. The slides from the resection specimens were viewed for the presence of vital tumor cells near the oral, ab-oral and circumferential resection margins, tumor type, tumor differentiation grade, extension of the tumor in the esophageal wall, and number of involved lymph nodes. R0 was defined as histologically tumor-free resection margins with ≥ 1 mm distance between tumor and resection margins. The 6th edition of the International Union Against Cancer (UICC) was used for TNM-classification, tumor grade, and stage grouping [14]. The grading of the therapy response was performed according to Junker et al. as described before [2, 13]. The degree of histomorphological regression in the esophageal wall, i.e. the effect of chemoradiation, was classified into four categories: tumor regression grade (TRG) I: more than 50% vital residual tumor cells; TRG II: 10-50% vital residual tumor; TRG III: less than 10% vital residual tumor cells; TRG IV: complete tumor regression, no evidence of vital tumor cells (Figure 1). Special attention was paid to therapy effects, such as mucin pools and squamous remnants in resection margins and lymph nodes (Figure 1). Patient characteristics are described using tables for categorical data, medians and ranges for continuous variables. Association between clinical and pathological param- eters was evaluated using the chi-square test. Overall survival time was calculated from the time of start of the chemoradiotherapy to the date of death or last follow-up. November 15 th 2007 was the censoring date for survival. Survival curves were estimated using the Kaplan-Meier method, and the log-rank test was used to evaluate the statistical significance of differences. The SPSS statistical package (version 12.0; SPSS Inc., Chicago, IL) was used for all statistical analyses. The 67 patients who were assessed in the current study had a median age of 59 years and the majority was male. Baseline patient characteristics are summarized in Table 1. The morphological changes in the tumor and in non-neoplastic tissue: such as mucin pools, squamous remnants, and stromal changes are shown in Figure 1. In Table 2 pathological characteristics of the resection specimens are summarized. A complete microscopic and macroscopic tumor resection was achieved in 66 of 67 patients (98.5%). In these 66 cases there were no mucin pools or areas containing keratin pearls or squamous remnants in the resection margins. When there were still vital tumor cells left in the resection specimen, the tumor differentiation grade was poorer compared to the tumor differentiation grade in the esophageal biopsies before chemoradiotherapy: Seven patients had differentiation grade 3/4 before chemoradiotherapy versus 27 patients after chemoradiation (p<0.001). TRG IV was achieved in 16 patients (24%), in 10 resection specimens (15%), little or no regressive changes were seen (TRG I). The response to chemoradiotherapy (as defined by TRG III and IV) was significantly higher in patients with squamous cell carcinoma (SCC) than in patients with adenocarcinoma (AC) (68% versus 54%, p=0.001). Poorly differentiated tumors (grade 3/4, esophageal biopsies) achieved significantly more tumor regression grade IV than good or moderately differentiated tumors (53% vs. 15%, p<0.01). Fifty-seven patients (85%) had tumor infiltration into the adventitia (T3) at baseline assessed by EUS, versus 25 patients (37%) after therapy as assessed by pathological analysis of the resection specimen, indicating significant downstaging by chemoradio- therapy (p<0.0001). Likewise, 38 patients (57%) had pre-treatment nodal involvement, as assessed by EUS, versus 21 patients (31%) post-therapy, as assessed by pathological examination of the resected specimen (p=0.006). In 11 of the 29 resected specimens with no pre-treatment nodal involvement (38%) regression effects were identified in the lymph nodes, suggesting downstaging from clinical stage N1 to pathological stage N0. The tumor regression grade of the primary tumor was significantly related to the risk of residual tumor cells in the esophageal wall: pT3 was found in 80% for TRG I, 65% for TRG II, and 25% for TRG III (p=0.01). This also applied for the resected lymph nodes: positive lymph nodes were found in 70% for TRG I, 50% for TRG II, 20% for TRG III, and 14% for TRG IV (p=0.008). After a median follow-up of 35 months (range 2-75), 20 patients relapsed (1 patient had a local/regional relapse, 9 patients developed distant metastases, and 10 patient developed both). The presence of pre-treatment nodal involvement was significantly associated with the development of distant metastases (14% vs. 42%, p=0.02). The median overall survival as calculated by the Kaplan Meier method has not yet been reached (Figure 2A). There was a trend towards a better overall survival in patients with a complete regression, i.e. TRG IV (p=0.09). The median overall survival for patients with TRG I- III was 57 months; the median survival for patients with TRG IV has not yet been reached (Figure 2B). Survival was significantly better in patients who had no nodal involvement before the start of treatment (median survival has not yet been reached) compared to patients who had nodal involvement as assessed by EUS (median survival 39 months, p=0.03) (Figure 3). No difference in overall survival could be found between patients with pathological nodal involvement compared to patients with no pathological nodal involvement, including 11 patients (16.7%) with therapy effects within the lymph nodes. There was no survival difference between patients with AC or SCC. In this study we described the histopathological changes seen after neoadjuvant chemoradiotherapy with paclitaxel and carboplatin in 67 cases of esophageal carcinoma. When vital tumor cells were left in the esophageal wall after chemoradiotherapy, the tumor was graded more frequently as poorly differentiated than before treatment. There are 3 explanations for the worsened differentiation grade in the resected specimens. First, poorly differentiated tumor cells may be more resistant to the cytotoxic insult caused ...

Context 5

... this resulted in a prolonged survival for these patients. Squamous cell carcinoma responded significantly higher to chemoradiotherapy than adenocarcinoma. This was not associated with a survival benefit. Neoadjuvant chemoradiotherapy led to a significant down-staging. Patients with pre-treatment nodal involvement had a significantly worse survival compared to patients with no pretreatment nodal involvement. This applied not for post-treatment nodal involvement. Conclusions. Poorly differentiated esophageal tumors significantly more frequent reach complete tumor regression after neoadjuvant chemoradiotherapy. The presence of nodal involvement as assessed by endoscopic ultrasound is significantly associated with a worse survival. Neoadjuvant chemoradiotherapy for patients with resectable esophageal cancer is increasingly used in an attempt to improve these patients. After chemoradiotherapy several histological changes in the resected specimen can be identified. These changes include cytological alterations such as cytoplasmic vacuolation and/or eosinophilia, nuclear pleomorphism, and necrosis, as well as stromal changes such as fibrosis (with or without inflammatory infiltrate), including giant cell granuloma formulation around ghost cells and mucin pools or keratin pearls at the site of the previous tumor [1-4]. On the basis of these changes, the grade of tumor regression can be defined. Regression of the primary tumor can range from the absence of regressive changes to a total response with no vital residual tumor cells. Several morphologic criteria have been defined for various types of cancer to objectively evaluate pathological response to neoadjuvant chemoradiation [1, 2, 5]. The prognosis of resectable esophageal cancer is mainly determined by the achieve- ment of a complete tumor resection (R0 resection), the depth of tumor infiltration, and the presence of lymph node metastasis [6-9]. Generally, the overall survival after neoadjuvant chemoradiotherapy is better in patients with a complete pathological response compared to patients with residual tumor cells [6, 8, 10-12]. Schneider et al. and Brücher et al. found that patients with only a few residual tumor cells within the resection specimen (<10%) had a significant better survival compared with patients with >10% residual tumor cells. In this article we review our experience with neoadjuvant carboplatin and paclitaxel concurrent with radiotherapy in patients with resectable esophageal cancer. We determined the residual tumor cells after preoperative treatment, and correlated the effect of specific pathologic and clinical findings to overall survival, in order to obtain prognostic information. We examined the esophageal biopsies and surgical specimens of all patients who were treated with neoadjuvant chemoradiation followed by surgery for esophageal cancer at the Erasmus Medical Centre between February 2001 and August 2006. Patients with T1N1 or T2-3N0-1 esophageal cancer were treated at our institution with five weekly courses of paclitaxel and carboplatin with concurrent radiotherapy followed by surgery in a phase II or phase III trial as described elsewhere [13]. The clinical data collected included patient age, sex, and preoperative clinical stage (obtained from CT scan and EUS). All haematoxylin and eosin (H&E) slides from preoperative tumor biopsies and from the resection specimens were evaluated by a specialized pathologist (H.v.D.). The preoperative tumor biopsies were viewed for histological tumor type, and grade. The resection specimen was evaluated using a standard protocol, providing information on margins, tumor type, tumor differentiation grade, extension of the tumor, and lymph nodes. If no macroscopically identifiable tumor was present, lesions such as an ulcer or an irregular area covered by mucosa were embedded in total together with surrounding areas in order to adequately judge the presence of residual tumor and therapy effects. The slides from the resection specimens were viewed for the presence of vital tumor cells near the oral, ab-oral and circumferential resection margins, tumor type, tumor differentiation grade, extension of the tumor in the esophageal wall, and number of involved lymph nodes. R0 was defined as histologically tumor-free resection margins with ≥ 1 mm distance between tumor and resection margins. The 6th edition of the International Union Against Cancer (UICC) was used for TNM-classification, tumor grade, and stage grouping [14]. The grading of the therapy response was performed according to Junker et al. as described before [2, 13]. The degree of histomorphological regression in the esophageal wall, i.e. the effect of chemoradiation, was classified into four categories: tumor regression grade (TRG) I: more than 50% vital residual tumor cells; TRG II: 10-50% vital residual tumor; TRG III: less than 10% vital residual tumor cells; TRG IV: complete tumor regression, no evidence of vital tumor cells (Figure 1). Special attention was paid to therapy effects, such as mucin pools and squamous remnants in resection margins and lymph nodes (Figure 1). Patient characteristics are described using tables for categorical data, medians and ranges for continuous variables. Association between clinical and pathological param- eters was evaluated using the chi-square test. Overall survival time was calculated from the time of start of the chemoradiotherapy to the date of death or last follow-up. November 15 th 2007 was the censoring date for survival. Survival curves were estimated using the Kaplan-Meier method, and the log-rank test was used to evaluate the statistical significance of differences. The SPSS statistical package (version 12.0; SPSS Inc., Chicago, IL) was used for all statistical analyses. The 67 patients who were assessed in the current study had a median age of 59 years and the majority was male. Baseline patient characteristics are summarized in Table 1. The morphological changes in the tumor and in non-neoplastic tissue: such as mucin pools, squamous remnants, and stromal changes are shown in Figure 1. In Table 2 pathological characteristics of the resection specimens are summarized. A complete microscopic and macroscopic tumor resection was achieved in 66 of 67 patients (98.5%). In these 66 cases there were no mucin pools or areas containing keratin pearls or squamous remnants in the resection margins. When there were still vital tumor cells left in the resection specimen, the tumor differentiation grade was poorer compared to the tumor differentiation grade in the esophageal biopsies before chemoradiotherapy: Seven patients had differentiation grade 3/4 before chemoradiotherapy versus 27 patients after chemoradiation (p<0.001). TRG IV was achieved in 16 patients (24%), in 10 resection specimens (15%), little or no regressive changes were seen (TRG I). The response to chemoradiotherapy (as defined by TRG III and IV) was significantly higher in patients with squamous cell carcinoma (SCC) than in patients with adenocarcinoma (AC) (68% versus 54%, p=0.001). Poorly differentiated tumors (grade 3/4, esophageal biopsies) achieved significantly more tumor regression grade IV than good or moderately differentiated tumors (53% vs. 15%, p<0.01). Fifty-seven patients (85%) had tumor infiltration into the adventitia (T3) at baseline assessed by EUS, versus 25 patients (37%) after therapy as assessed by pathological analysis of the resection specimen, indicating significant downstaging by chemoradio- therapy (p<0.0001). Likewise, 38 patients (57%) had pre-treatment nodal involvement, as assessed by EUS, versus 21 patients (31%) post-therapy, as assessed by pathological examination of the resected specimen (p=0.006). In 11 of the 29 resected specimens with no pre-treatment nodal involvement (38%) regression effects were identified in the lymph nodes, suggesting downstaging from clinical stage N1 to pathological stage N0. The tumor regression grade of the primary tumor was significantly related to the risk of residual tumor cells in the esophageal wall: pT3 was found in 80% for TRG I, 65% for TRG II, and 25% for TRG III (p=0.01). This also applied for the resected lymph nodes: positive lymph nodes were found in 70% for TRG I, 50% for TRG II, 20% for TRG III, and 14% for TRG IV (p=0.008). After a median follow-up of 35 months (range 2-75), 20 patients relapsed (1 patient had a local/regional relapse, 9 patients developed distant metastases, and 10 patient developed both). The presence of pre-treatment nodal involvement was significantly associated with the development of distant metastases (14% vs. 42%, p=0.02). The median overall survival as calculated by the Kaplan Meier method has not yet been reached (Figure 2A). There was a trend towards a better overall survival in patients with a complete regression, i.e. TRG IV (p=0.09). The median overall survival for patients with TRG I- III was 57 months; the median survival for patients with TRG IV has not yet been reached (Figure 2B). Survival was significantly better in patients who had no nodal involvement before the start of treatment (median survival has not yet been reached) compared to patients who had nodal involvement as assessed by EUS (median survival 39 months, p=0.03) (Figure 3). No difference in overall survival could be found between patients with pathological nodal involvement compared to patients with no pathological nodal involvement, including 11 patients (16.7%) with therapy effects within the lymph nodes. There was no survival difference between patients with AC or SCC. In this study we described the histopathological changes seen after neoadjuvant chemoradiotherapy with paclitaxel and carboplatin in 67 cases of esophageal carcinoma. When vital tumor cells were left in the esophageal wall after chemoradiotherapy, the tumor was graded more frequently as poorly differentiated than before treatment. There are 3 explanations for the worsened ...