Preoperative mitomycin, ifosfamide, and cisplatin followed by esophagectomy in squamous cell carcinoma of the esophagus: pathologic complete response induced by chemotherapy leads to long-term survival.

PURPOSE: Squamous cell carcinoma of the esophagus remains an aggressive disease with a poor prognosis, even after curative-intent surgery. This article analyzes the impact of preoperative chemotherapy with mitomycin, ifosfamide, and cisplatin (MIC) on a cohort of 68 patients. PATIENTS AND METHODS: From 1988 to 1994, 68 patients with potentially operable squamous cell carcinoma of the esophagus were entered onto two phase II trials of neoadjuvant chemotherapy with mitomycin 6 mg/m2, ifosfamide 3 g/m2, and cisplatin 50 mg/m2 and received between two and four cycles of treatment at 3-weekly intervals. Two patients were removed from the analysis when they were found to have malignancy other than squamous cell carcinoma of the esophagus. RESULTS: Forty (61%) of 66 patients had a radiologic response to chemotherapy (18 complete responses and 22 partial responses), and 52 (79%) of 66 patients went on to have the primary tumor resected. There were nine pathologic complete responders, seven of whom remain fit and well after at least 60 months of follow-up. The overall median survival was 12.4 months (95% confidence interval, 9.6 to 18.8 months). The complete response and node-negative patients survived significantly longer than those in other categories (log-rank chi2 = 18.8; P <.001): on average 13 months longer than the node-positive or nonresected category (22.0 v 9.4 months). The toxicity of the regimen was low. CONCLUSION: MIC is an easily administered, well-tolerated, and efficacious regimen as neoadjuvant therapy for patients with squamous cell carcinoma of the esophagus. These results warrant further investigation.

Effects of chemotherapy on ultrastructure of oesophageal adenocarcinoma.

AIMS: To compare and contrast the ultrastructural appearance of resected oesophageal adenocarcinomas treated with preoperative chemotherapy with that of non-treated resected controls; and to determine the usefulness of this method in the assessment of the effectiveness of the chemotherapeutic regimen. METHODS: Ten resected oesophageal adenocarcinomas treated with preoperative chemotherapy--mitomycin-C, ifosfamide, and cisplatin (MIC)--were examined by transmission electron microscopy and their appearance compared with that of 13 concurrent untreated resected oesophageal adenocarcinomas. RESULTS: The treated adenocarcinomas showed cytotoxic damage although complete tumour eradication was not achieved. In all 10 treated cases a variable proportion of the neoplastic cells showed unusual degenerative and necrotic changes not seen in untreated cases. In the most affected carcinomas the stroma contained increased numbers of inflammatory cells. CONCLUSIONS: This ultrastructural method is useful for the assessment of the in vivo effect of MIC.

Expression of HLA-ABC, HLA-DR and intercellular adhesion molecule-1 in oesophageal carcinoma.

AIM: To examine the expression of HLA-ABC and HLA-DR major histocompatibility (MHC) antigens and intercellular adhesion molecule (ICAM)-1 in normal, inflamed, metaplastic, and neoplastic oesophageal tissue and in freshly disaggregated tumours. METHODS: Sequential sections of frozen tissue and cytospins of freshly disaggregated tumour were stained using the ABC peroxidase system and monoclonal antibodies specific for HLA-ABC, HLA-DR and ICAM-1. RESULTS: Normal oesophageal tissue showed positive staining for HLA-ABC in the basal layers of the oesophageal squamous epithelium and on the epithelial cells of the submucosal oesophageal glands. HLA-DR and ICAM-1 were not detected in either of these cell types. In 20 of 37 (54%) carcinomas HLA-ABC was expressed weakly, with heterogeneous expression in nine (24%). Two tumours showed strong expression of HLA-ABC, but 15 of 37 (41%) were negative. HLA-DR and ICAM-1 were expressed weakly in six of 37 (16%) carcinomas without correlation with each other or with the expression of HLA-ABC. CONCLUSIONS: HLA-ABC is absent from a high proportion of oesophageal carcinomas (41%) and is otherwise variably and weakly expressed with strong expression in only a small fraction (3%). In other carcinomas there is a higher level of HLA-ABC expression. This discrepancy may partly explain the aggressive nature of oesophageal carcinomas. HLA-DR and ICAM-1 are not normally expressed on those cells from which oesophageal carcinomas are thought to arise. The limited expression found here could suggest a partial or inhibited immune response against oesophageal carcinoma. In vivo repressive factors may be involved.

Lack of correlation of P-glycoprotein expression with response to MIC chemotherapy in oesophageal cancer.

The multidrug resistance gene product P-glycoprotein (P-GP) was assessed immunohistochemically (by antibody JSB-1) in biopsy specimens from 27 oesophageal squamous carcinomas and 10 adenocarcinomas before treatment with mitomycin, ifosfamide and cisplatin (MIC). Tumours were assessed following treatment and correlation with response sought. Of the squamous carcinomas, 74% (20/27) responded to MIC but only one expressed P-GP before and after treatment. Of the adenocarcinomas, 30% (three of 10) responded. Seven of the 10 adenocarcinomas expressed P-GP before treatment but all 10 were P-GP positive after chemotherapy. The difference in prevalence and induction of P-GP between the histological types was highly significant and may correlate with the greater response to MIC seen in squamous carcinomas compared with adenocarcinomas. P-GP cannot be used as a predictive marker of response as tumours express it inconsistently with response to MIC. Resistance to MIC may be due to other mechanisms.

A case of systemic pseudo-pseudoxanthoma elasticum with diverse symptomatology caused by long-term penicillamine use.

A 47 year old man presented with a two year history of increasing cervical dysphagia, dyspnoea, and cutaneous signs. He had been diagnosed 27 years previously with Wilson's disease and was treated with penicillamine (1.5 g daily). Systemic abnormality of elastic fibres was confirmed by light and electron microscopy following biopsy of skin, lung, oesophageal muscle, gum, pharyngeal tissue, and cervical connective tissue. Dysphagia was relieved by cricopharyngeal myotomy. Substitution of trientene dihydrochloride for penicillamine relieved cutaneous and systemic manifestations. This is possibly the first case demonstrating an association between prolonged penicillamine use and biopsy proved systemic pseudo-pseudoxanthoma elasticum. The presenting symptoms may have resulted from the abnormal numbers and properties of elastic fibres, and the changes were caused by penicillamine use, rather than by idiopathic, inherited pseudoxanthoma elasticum.

A primary tumour of the oesophagus with both melanocytic and schwannian differentiation. Melanocytic schwannoma or malignant melanoma?

A 76 year old white woman presented with a four month history of dysphagia and weight loss. Clinical, radiological, and endoscopic examination revealed a pigmented mass in the lower third of the oesophagus. The preoperative diagnosis, including biopsy examination, was that of malignant melanoma. Following oesophageal resection, the mass was found to be a localised, relatively superficial tumour with light, electron microscopic, and immunohistochemical features common to both Schwann cells and melanocytes. The patient survived 46 months after surgery and died of a stroke, with no evidence of tumour recurrence. The tumour is presented as a case of melanocytic schwannoma, with unique features when compared with oesophageal melanotic schwannomas and malignant melanomas described in the literature. The differential diagnosis is discussed and an origin from a common precursor cell of neural crest origin is postulated.

Histopathological findings in oesophageal carcinoma with and without preoperative chemotherapy.

AIMS: To investigate the pathological effects of preoperative chemotherapy on oesophageal carcinoma. METHODS: Qualitative and quantitative changes in oesophageal carcinoma after preoperative chemotherapy were assessed by examination of biopsy specimens before treatment and resected specimens. RESULTS: Of 13 patients with adenocarcinoma treated with 5-fluorouracil, adriamycin, and mitomycin (FAM), nine showed minor histological changes compared with 14 control cases. All 12 patients with squamous carcinoma treated with preoperative mitomycin, ifosfamide, and cisplatin (MIC) showed noticeable histological changes when compared with the 13 control cases. Changes included complete ablation (n = 1) and partial regression (n = 5) of the tumour. A quantitative estimate of the proportion of tumour to stroma showed no difference between control adenocarcinomas and those treated with chemotherapy. There was, however, a significant reduction (p < 0.01) in the proportion of tumour to stroma in the treated squamous group compared with the controls. There was no relation between the degree of response in squamous carcinomas and the degree of differentiation of the tumour. Patients in which squamous carcinomas responded well, as assessed quantitatively, showed a tendency to better survival at one year. CONCLUSIONS: Histopathological changes attributable to chemotherapy can be observed in oesophageal carcinoma. The response of squamous carcinoma to MIC is histologically more evident than that of adenocarcinoma to FAM. A quantitative technique may be useful in assessing the effect of chemotherapy in oesophageal squamous carcinoma.

Expression of E-cadherin in oesophageal carcinomas from the UK and China: disparities in prognostic significance.

AIMS: To study the expression and prognostic significance of the cell adhesion molecule E-cadherin in oesophageal tumours from the UK (low risk area) and China (high risk area). METHODS: E-cadherin expression was measured immunohistochemically in resected tumours from 17 patients in the UK with adenocarcinoma, 23 patients from the UK with squamous carcinoma, and 30 patients from China with squamous carcinomas who survived for five years postoperatively and compared with similar tumours from patients in the same regions who did not survive (140 tumours in all). RESULTS: Normal squamous epithelial cells and well differentiated areas of tumours showed membranous staining for E-cadherin expression. Cytoplasmic staining, heterogeneous staining, or an absence of staining was seen in dysplastic epithelium and in less well differentiated areas of tumours. Only one of 140 primary tumours had homogeneous membranous expression. In tumours from UK patients with adenocarcinoma (p = 1.00) and from Chinese patients with squamous carcinomas (p = 0.06) there was no correlation between E-cadherin absence and non-survival. In tumours from UK patients with squamous carcinomas there was a significant correlation between absence of E-cadherin and non-survival (p = 0.009). Tumours from UK patients with squamous carcinoma who survived were significantly less likely to be E-cadherin absent than those from Chinese patients with squamous carcinomas who survived (p = 0.007). Multivariate analysis (n = 37 UK, paired data) showed that absence of E-cadherin in the primary tumour was a weak independent prognostic factor for non-survival (30% significance level; p = 0.26; odds ratio = 3.56). In UK nodal metastases there was no correlation between E-cadherin expression and survival. CONCLUSIONS: Squamous carcinomas from UK patients differed from both adenocarcinomas from UK patients and carcinomas from Chinese patients with respect to E-cadherin expression and prognostic significance. In tumours from UK patients, E-cadherin absence in the primary carcinoma (a weak independent prognostic factor) but not metastases correlated with non-survival.

Lymphocyte infiltration in oesophageal carcinoma: lack of correlation with MHC antigens, ICAM-1, and tumour stage and grade.

Infiltration by T lymphocytes into oesophageal carcinomas was assessed immunohistochemically, total T lymphocyte numbers by staining for CD3 and activated T lymphocytes by staining for CD25. Five squamous carcinomas and seven adenocarcinomas, resected without neoadjuvant treatment, were studied. Computer aided quantitation showed that total numbers of tumour infiltrating CD3 positive cells were highly variable (range 48-1673 cells/mm2). They were located largely in the stromal (87.9-99.2%) rather than intratumoral regions. Up to 84% of tumour infiltrating T lymphocytes were CD25 positive, although the median figure was 33%. There was no correlation between T lymphocyte infiltration or activation and expression of class I and II histocompatibility antigens, intercellular adhesion molecule-1, tumour stage or grade. These results imply that the local inflammatory response in oesophageal carcinomas is deregulated, which may be a factor contributing to the aggressive nature of the tumours.

The metabolic marker tumour pyruvate kinase type M2 (tumour M2-PK) shows increased expression along the metaplasia-dysplasia-adenocarcinoma sequence in Barrett's oesophagus.

BACKGROUND: Proliferating and tumour cells express the glycolytic isoenzyme, pyruvate kinase type M2 (M2-PK). In tumours cells, M2-PK usually exists in dimeric form (tumour M2-PK), causing the accumulation of glycolytic phosphometabolites, which allows cells to invade areas with low oxygen and glucose concentrations. AIMS: To investigate the expression of tumour M2-PK during the metaplasia-dysplasia-adenocarcinoma sequence of Barrett's oesophagus, and to assess the prognostic usefulness of tumour M2-PK in oesophageal cancer. MATERIALS/METHODS: One hundred and ninety cases selected from the histopathology archives as follows: 17 reflux oesophagitis, 37 Barrett's oesophagus, 21 high grade dysplasia, 112 adenocarcinomas, and three control tumours. Sections were stained immunohistochemically with antibody to tumour M2-PK. RESULTS: Tumour M2-PK was expressed in all cases, and increased cytoplasmic expression was seen with progression along the metaplasia-dysplasia-adenocarcinoma sequence. All cases of adenocarcinoma showed 100% staining so that tumour M2-PK was not a useful prognostic marker. CONCLUSIONS: Tumour M2-PK is not a specific marker of Barrett's adenocarcinoma, but may be important as a marker of transformed and highly proliferating clones during progression along the metaplasia-dysplasia-adenocarcinoma sequence.

What is the optimal distal resection margin for esophageal carcinoma?

BACKGROUND: Whereas a proximal resection margin of 12 cm is recommended for complete resection of esophageal cancer, the extent of distal resection is unclear. METHODS: We examined distal resection margins in a consecutive series of patients who underwent esophagectomy for squamous cell carcinomas (n = 50), primary esophageal adenocarcinomas (n = 100), and adenocarcinomas of the cardia (n = 39), in whom all macroscopic tumor was judged to be completely resected. RESULTS: Microscopic tumor was found at a 3-cm distal resection margin for one multifocal squamous cell carcinoma. Positive distal resection margins were seen in 12% (12 of 100 patients) of primary esophageal adenocarcinomas (median, 2 cm versus 4 cm if negative; p = 0.002, Wilcoxon) and 28% (11 of 39 patients) of cardia adenocarcinomas (median, 1 cm versus 3 cm if negative; p = 0.02, Wilcoxon). Although pathologic stage was shown to be the only significant predictor of overall survival (Hazard ratio [HR] 1.8; 95% confidence interval 1.2 to 2.6; p = 0.007), there was a trend toward reduced postoperative survival for patients with histologically positive distal resection margins, in particular for patients with cardia adenocarcinomas (median, 15.4 months versus 5.7 months if negative; p = 0.0001). CONCLUSIONS: To achieve consistently negative distal resection margins, we recommend resection of at least 5 cm of macroscopically normal foregut below the distal margin of the primary tumor.

Phase II study of mitomycin, ifosfamide and cisplatin in adenocarcinoma of the oesophagus.

We evaluated the effect of brief neoadjuvant chemotherapy in patients with apparently operable adenocarcinoma of the oesophagus. Two courses of mitomycin (6 mg/m2), ifosfamide (3 g/m2) and cisplatin (50 mg/m2;MIC) were given followed by evaluation of response by barium swallow and computed tomography scan. Of 20 patients, 17 completed both courses and 4 (20%) showed a partial response. Toxicity was generally mild and consisted principally of nausea and vomiting. Altogether, 15 patients were surgically explored; resection was completed in 12 patients, 3 of whom died in hospital (25%). Neoadjuvant therapy with MIC offers no advantage over surgery alone.

Weight gain as an indicator of response to chemotherapy for oesophageal carcinoma.

Patients with oesophageal carcinoma commonly present with dysphagia and weight loss, which may be related to the tumour burden and/or the physical obstruction to the passage of food. In this study we have examined the relationship between weight change and response to chemotherapy in 28 patients undergoing neo-adjuvant chemotherapy for squamous or anaplastic carcinoma. Two pulses of mitomycin, ifosfamide and cisplatin were given 3 weeks apart. Body weights were measured prior to the first pulse and 3 weeks after the second. Patients underwent oesophageal dilatation routinely at diagnostic endoscopy prior to chemotherapy, in order to permit oral nutrition. No dietary modifications were made. Tumour response was assessed on a barium swallow. Patients had a normal spread of weights on presentation. In the non-responding group (n = 9), eight patients lost weight and one gained weight. Of the partial responders (> 50% tumour shrinkage; n = 11), five gained weight, five lost weight and one remained constant. In the complete response group (n = 8), six gained weight and two lost weight. Statistical analysis showed a significant difference (F = 4.61; P = 0.02) between change in weight expressed as a percentage of ideal weight in nonresponders (mean -5.3%) versus partial responders (mean +2.4%), and in non-responders versus complete responders (mean +1.1%). Weight gain during chemotherapy is a good indication of response, although its absence does not preclude a response. In the majority of patients who respond to chemotherapy there will be an increase in weight with improvement in their general condition prior to operation.

Combined results from three phase II trials of neoadjuvant chemotherapy in operable adenocarcinoma of the oesophagus.

Adenocarcinoma of the oesophagus is a systemic disease at presentation in the majority of patients. This article analyses the impact of preoperative chemotherapy on a cohort of 68 patients. From 1990 to 1996, 68 patients with potentially operable adenocarcinoma of the oesophagus were entered into three sequential Phase II trials of neoadjuvant chemotherapy with cisplatin/mitomycin C/ifosfamide, cisplatin/5-fluorouracil (5-FU) and mitomycin C/cisplatin/5-FU. Twenty-four (35%) patients had a radiological (4 complete; 20 partial) response to chemotherapy, and 52 (76%) went on to have the primary tumour resected. There was only one pathological complete responder. The overall median survival was 13 months (95% confidence interval (CI) 9-16). Survival for the 28 N(0) patients was 34 months (95% CI 14-60). The pattern of failure for resected patients was predominantly systemic (16/17). These results indicate that neoadjuvant chemotherapy followed by surgery for adenocarcinoma of the oesophagus achieves excellent local control. The dominance, however, of distant recurrence after surgery underlines the fact that, in the majority of patients, the only hope of improving results in the future is to develop better systemic therapies.

A phase II trial of cisplatin and 5-fluorouracil in adenocarcinoma of the oesophagus.

The effect of brief neoadjuvant chemotherapy in patients with apparently operable adenocarcinoma of the oesophagus has been investigated. Two courses of cisplatin and 5-fluorouracil (CFu) were given, followed by evaluation of the response by barium swallow. Twenty-one of 23 patients completed both courses. Two showed a complete response and five a partial response. In only one patient was there a pathological complete response. Toxicity was mild and consisted principally of nausea and vomiting. All patients underwent surgical exploration; resection was completed in 17. There were three hospital deaths (18%). Although CFu has produced two complete responses (on barium swallow) and one complete pathological clearance of tumour, the disappointing total response rate of 7/21 (33%; 95% CI 13-53) or 7/23 (30%; 95% CI 12-49) leads us to believe that further Phase II trials are needed to identify more efficacious agents and regimens.

A phase II trial of preoperative mitomycin, cisplatin and 5-fluorouracil in adenocarcinoma of the oesophagus.

The effect of neoadjuvant chemotherapy in patients with apparently operable adenocarcinoma of the oesophagus has been investigated. Two courses of mitomycin, cisplatin and 5-fluorouracil (MCF) were given, followed by a radiological evaluation of response. Twenty-two of 25 patients completed both courses. Two showed a complete response and 12 a partial response. There was a pathological complete response of the primary tumour in only one patient (although there was residual secondary tumour in a local lymph node). The main toxicity was myelosuppression, with 9/22 patients having the second chemotherapy course delayed. There were three sudden deaths, one due to a pulmonary embolus and two due to complications of infections. Twenty-one patients underwent surgical exploration; there were 18 resections. Although the radiological response rate of MCF (14/25; 56%; 95% CI 37-75) appeared promising, there were no pathological complete responders. Further Phase II trials are needed to identify more efficacious agents and regimens that will yield a pathological response rate of at least 10%, before proceeding to randomized trials of neoadjuvant chemotherapy for adenocarcinoma of the oesophagus.

A phase II trial of four courses of preoperative chemotherapy in squamous or anaplastic carcinoma of the oesophagus.

We have reported the results of a previous Phase II trial of two courses of neoadjuvant mitomycin (6 mg/m2), ifosfamide (3 g/m2) and cisplatin (50 mg/m2) (MIC) in squamous or anaplastic carcinoma of the oesophagus. In this current study, we have investigated whether there was any clinical benefit in extending the preoperative treatment to four courses for patients who responded after two courses. Response was assessed by barium swallow, which was compared with previous barium swallows performed prior to any treatment and after the second course of MIC. Of an initial 43 patients, 27 (63%) were assessed as responders after two courses of MIC. Twenty of these 27 patients were entered into the study with a view to receiving two further courses of MIC prior to surgery. Seventeen completed four courses. Five patients were complete responders after two courses and remained complete responders after four courses. Twelve patients were partial responders after two courses; six of these became complete responders after four courses, five remained partial responders, and one showed progression. Haematological toxicity and alopecia were increased after extending the number of courses beyond two. On pathological assessment, three patients with a complete response after four courses, and one with a complete response after three courses, had microscopic clearance of tumour. Extension beyond two courses of neoadjuvant MIC gives an improvement in response, as judged by barium assessment, but increases toxicity, cost of treatment and delay before surgery. Although the numbers are small, the results suggest that a worthwhile improvement in the radiological response of squamous or anaplastic oesophageal tumours may be gained by proceeding beyond two courses of MIC. A randomized trial, with larger numbers of patients, is needed to show whether there is any improvement in radiological and pathological response rates and in survival to be gained by the extension of treatment beyond two courses.

A phase II study of mitomycin, ifosfamide and cisplatin in operable and inoperable squamous cell carcinoma of the oesophagus.

We have evaluated the effect of mitomycin 6 mg/m2, ifosfamide 3 g/m2, and cisplatin 50 mg/m2 (MIC) in two groups of patients with squamous or undifferentiated carcinoma of the oesophagus, as either preoperative or primary treatment. Response was assessed by barium swallow, CT scan, and measurement of metastases where present. Toxicity was acceptable and there were no chemotherapy related deaths. In the operated group, five of 23 patients (22%) showed a complete response (three confirmed histologically) and nine (39%) showed a partial response following two courses of MIC. Resection was completed in 21 patients, with three hospital deaths (14%). Of the 18 patients who were discharged from hospital, eight have died at 4-24 months (median 13) from the start of treatment and 10 are alive at 5-35 months, with known recurrence in one. In the non-operated group, five of 20 patients (25%) showed a response, one complete, following one to four (mean 2.6) courses of MIC. Nineteen patients have died (at median 5 months), and one, who had a complete response, is alive and free from disease at 29 months. Neo-adjuvant therapy with MIC in squamous carcinoma of the oesophagus has shown encouraging early results, with acceptable toxicity.

Five newly established oesophageal carcinoma cell lines: phenotypic and immunological characterization.

The derivation of permanent cell lines from 40 resected oesophageal carcinomas has been attempted. Five long-term lines have been established from three adenocarcinomas, one mixed carcinoma and one squamous carcinoma. Molecular and cellular analyses have been carried out on the lines and clones derived from them. Karyotype analysis indicates genetic variation among the clones. HLA-A, -B and -C is expressed constitutively, but not HLA-DR. ICAM-1-expressing phenotypes may have arisen during adaptation to long-term culture. All lines are capable of response to interferon-gamma (IFN-gamma) and all produce transforming growth factor beta 1 (TGF-beta 1). Two lines are resistant to the inhibitory growth effects of the latter, possibly contributing to malignancy. It is anticipated that these lines, originating from histologically different carcinomas, will provide a valuable, continuous resource for the investigation and treatment of these aggressive tumours.

Tumour heterogeneity: a problem in biopsy assessment of the proliferation index of oesophageal adenocarcinomas.

Tumour heterogeneity may pose a problem when biopsy specimens are taken to measure proliferation (for example, in assessing response to therapy). Two "biopsy specimens" were taken from the centre and two from the edge of the luminal surface of 20 resected oesophageal adenocarcinomas. The proliferation index for each "biopsy specimen" was measured by counting Ki67 labelled nuclei in histological sections. The proliferation index was not associated with tumour differentiation or stage. There was site specific heterogeneity with a significant difference in proliferation index between the central (mean (SD) 36.4 (9.7)) and edge "biopsy specimens" (39.3 (9.9)). There was, however, a wide range of differences between pairs of "biopsy specimens" from both sites. In conclusion, if a tumour is to be sampled for measurement of the proliferation index before and after treatment, then the sequential biopsy specimens (preferably duplicated on each occasion) should be taken consistently from a leading edge of the lesion.