Journal of Cancer Research and Therapeutics Medknow Publications on behalf of the Association of Radiation Oncologists of India (AROI) ISSN: 0973-1482 EISSN: 1998-4138 Vol. 3, Num. 1, 2007, pp. 2-7

Journal of Cancer Research and Therapeutics, Vol. 3, No. 1, January-March, 2007, pp. 2-7

Original Articles - Clinical

Factors influencing the development of ulcers and strictures in carcinoma of the esophagus treated with radiotherapy with or without concurrent chemotherapy

Khurana Rohini, Dimri Kislay, Lal Punita, Rastogi Neeraj, Joseph K, Das Maria, Kumar Shaleen

Department of Radiotherapy, Sanjay Gandhi Postgraduate Institute of Medical Sciences, Raebareli Road, Lucknow - 226 014
Correspondence Address:Department of Radiotherapy, Sanjay Gandhi Postgraduate Institute of Medical Sciences, Raebareli Road, Lucknow - 226 014, shaleenkr@yahoo.com

Code Number: cr07002

Abstract

Keywords: Brachytherapy, chemotherapy, esophageal neoplasms, late morbidity, radiotherapy

Introduction

Materials and Methods

Between September 1990 and December 2005, patients of cancer esophagus receiving definitive treatment in our institution were identified from our database. The inclusion criteria were previously untreated patients of non-metastatic, inoperable squamous cell carcinoma of esophagus with Karnofsky performance score (KPS) of at least 50. Following a detailed history that included grade and duration of dysphagia and any prior interventions, patients underwent pretreatment evaluation, in the form of a barium swallow, endoscopy and biopsy along with a chest X-ray PA view, hemograms and liver and kidney function tests. Contrast enhanced computed tomography (CECT) of the neck, thorax and upper abdomen were more consistently advised to patients since 1996. Metastatic lymph nodes were defined if they were > 1 cm in greatest diameter on CECT while the descending aorta was judged to be involved if the contact angle with the tumor was 90 degrees or greater.

The patients were assembled into one of the following groups: those receiving at least 60Gy EBRT (Gp E); EBRT followed by HDR-ILRT (Gp E+I); at least 50Gy EBRT with chemotherapy (CT) (Gp E+C); EBRT+HDR-ILRT + CT (Gp E+I+C). EBRT was delivered with mega voltage equipment using anterior and posterior opposed fields up to 36 Gy/18 fraction/five days a week to primary tumor with a 5 cm margin cranio-caudally and a lateral margin of 2 cm. Any further doses (typically 14-30Gy) were given to a smaller, off-cord fields with 2 cm craniocaudal and 1 cm radial margins using one anterior and two posterior oblique fields. However for the patients receiving HDR-ILRT, the phase II EBRT dose was usually a maximum of additional 14 Gy/7 fractions/five days a week and ILRT was planned one to two weeks following completion of EBRT with or without concurrent CT. Lung heterogeneity corrections were made but no portal verification was performed.

HDR-ILRT was given as 2 fractions of 6 Gy each spaced a week apart with or without concurrent CT if the esophageal lumen was negotiable. ILRT was delivered using a HDR remote after loading system with a 6 mm diameter applicator using an Ir-192 source (Microselectron, Nucletron, Netherlands) or a 10 mm applicator diameter using a Co-60 source (Ralstron 20 B, Shimadzu, Japan). The dose was specified at 5 mm from the surface of the applicator. However, if the applicator could not pass, further EBRT dose was planned and ILRT was tried once again after 60 Gy of EBRT.

Concurrent CT was given as cisplatin 35 mg/m ² IV infusion weekly dose with adequate hydration and anti emesis and mannitol diuresis. Weekly hemogram and kidney function tests were advised and CT was postponed if the white blood cell count was < 3500/mm ³ but no dose reduction was done. Patients were examined weekly during treatment and acute toxicity assessment was done according to RTOG/EORTC criteria. Ryle's tube or inpatient management with parenteral nutrition according to the severity of dysphagia was advised during treatment.

After treatment completion, first follow-up evaluation was done at one month and thereafter at two months for the first year and at three months for the subsequent years. Besides a physical examination at each follow-up visit, a barium swallow was advised and if any irregularity or constriction seen, patients were advised for endoscopy with a note encouraging biopsy if any suspicious lesion was seen at endoscopy.

Patient were considered to be locally disease-free if barium swallow was smooth and there were no signs or symptoms of spread to mediastinum such as vocal cord palsy and a negative biopsy if performed. Endoscopic evaluation and biopsy were performed only to evaluate the cause of recurrent / persistent dysphagia or odynophagia and to relieve the symptoms by dilatation. Ulcers observed during endoscopy were biopsied and scored as treatment related if reported negative for malignant cells. Similarly strictures were reported as treatment related if the barium swallow was smooth and there were no malignant cells at biopsy. In the absence of biopsy proof, a smooth stricture when detected and followed by a survival of greater than 6 months, was considered as treatment induced. Late morbidity was scored only for those strictures causing significant morbidity - defined as requiring bougie dilatation - i.e., grade 3 RTOG / EORTC late morbidity criteria.[8]

Demography and intervention variables are stated as summary measures (mean, medium, standard duration (SD) and range) along with percentages and proportions. Patient, tumor or treatment related variables were compared using the c ² test, t-test or analysis of variance (ANOVA). Crude incidences of radiation induced ulcers and strictures are presented along with a univariate and multivariate analysis of factors likely to influence their development. The time to ulcer or stricture development was calculated from the time of initiation of treatment. Although not the purpose of this audit, to place the morbidity in perspective, overall survival was calculated using Kaplan-Meier method with patients dying of any cause and those lost to follow-up considered as events assuming the 'worst case scenario'. The log-rank test was used to ascertain significance of difference between survival curves. Analysis was carried out at of October 2006.

Results

There were 244 patients who could be assigned to the following treatment groups- Gp E, n=44; Gp E+I, n=98; Gp E+C, n=68 and Gp E+I+C, n=34. The patient and tumor characteristics are listed in [Table - 1]. All the treatment groups were matched for various variables however pretreatment hemoglobin and KPS score were lower in group E - the possible explanation being that the patients in lower KPS patient or those with lower hemoglobin levels were not considered for protocols that included CT. Also, the lower incidence of lymph node metastasis in group E+I reflected the lower frequency of CT scans during the period 1990-1996 before the inclusion of CT in the treatment protocols.

The treatment received by the various groups and development of morbidity is compared in [Table - 2]. The incidence of treatment induced ulcers was 7%, 8%, 6%, 21% ( P =0.08) while that of strictures was 14%, 9%, 21%, 41% ( P =0.00) for the groups E, E + I, E+C and E+I+C respectively. Median time to ulceration (in months) was - 3.4, 6, 6.2, 4.9 while median time to development of strictures (in months) was - 5.1, 4.8, 6.6, 5.9 respectively for the groups E, E+C, E+I and E+I+C respectively.

On univariate analysis [Table - 3], significant factors affecting stricture development were - HDR-ILRT with narrow applicators ( P =0.00), usage of concurrent chemotherapy ( P =0.00) those assigned to group E+I+C ( P =0.00) and those with better KPS ( P =0.03). For ulcers, there was a trend towards increase in patients treated with all the three modalities i.e. group E+I+C ( P =0.08). Logistic regression [Table - 4] retained only group E+I+C for development of ulcers (OR 10.36 95% CI 1.2 - 89.1, P =0.03) and strictures (OR 4.2, 95% CI 1.4-12.2 P =0.00).

The median survival in groups E, E+I, E+C and E+I+C were 9, 10, 11 and 14.5 months with the corresponding 5-year survival probability being 12%, 10%, 14% and 25% (Log-rank P = 0.326) [Figure - 1]. For groups E, E+I, E+C and E+I+C; 2, 9, 4 and 3 patients respectively were lost to follow up at varying periods and their fate unknown to us. They were considered dead (of unknown causes) for the purpose of computation of the survival curves.

Discussion

The RTOG 85-01 trial was the landmark trial that showed superiority of chemo-radiotherapy (CRT) over RT alone.[6] Following the unequivocal advantage of CRT as demonstrated by Herskovic et al , intensification of RT in the form of HDR-ILRT boost was tested in the RTOG 92-07 phase I/II study where an unacceptably high incidence of fistulas and treatment related deaths resulted in this strategy being abandoned.[7] This trial was perhaps one of the first reports that fuelled concerns of the combined morbidity of a CRT protocol that incorporated HDR-ILRT. In this audit, the survival expectations of the group E+I+C were the best compared to all other groups, but the primary aim was not to assess the efficacy of various interventions, as is obvious that fitter patients were assigned to receive more aggressive treatment protocols that included concurrent CT [Table - 1]. Only an adequately powered randomized trial can provide appropriate evidence to establish a survival advantage and our current focus was on dissecting out reasons for the increased morbidity seen in group E+I+C compared to any of the other groups.

The incidence of 6-8% ulcerations, 9-21% strictures in groups E, E+I and E+C is similar to that reported by Sur et al[3],[4] and the follow-up report of RTOG 85-01 where 20% grade three late esophageal toxicity was reported.[9] However, in the present audit, a three-fold increase in the incidence of ulcers and a two to four-fold increase in the incidence of strictures were observed in group E+I+C compared to the other groups. Direct comparison of morbidity between different trials is a difficult task because variables that could affect the development of late morbidity such as EBRT doses, interval between EBRT and HDR-ILRT, ILRT fraction sizes and ILRT applicator sizes vary simultaneously. For example, Hishikawa et al[10] in 148 patients used EBRT 60Gy/30fx/6weeks, followed a week later with 2 x 6Gy HDR-ILRT prescribed at 5 mm from surface of a 10 mm diameter applicator, reported 28%, 10% and 4% of ulcers, strictures and fistulae respectively. Gaspar et al[11] in 27 patients reported a 0% stricture rate and 3.7% fistula rate with EBRT 50Gy/25fx/5 weeks followed two to three weeks later with 2 x 8Gy HDR-ILRT prescribed at 8 mm from surface of a 4 mm diameter applicator. In the arm randomized to receive ILRT, Okawa et al[5] reported on 43 patients who received EBRT 60Gy/30fx/6 weeks followed by 2 x 5Gy HDR-ILRT (or low dose rate ILRT) prescribed at 5 mm from the surface of a 10 mm diameter applicator and documented a 5% ulceration and 2% stricture rate. In sharp contrast, Hama et al ,[12] using an identical dose schedule and prescription point (5 mm from surface) but with a 15 mm diameter balloon applicator reported a 50% fistula rate in six consecutive patients. Intuitively, the larger diameter applicator with a reduced hyperdose sleeve[13] cannot explain this morbidity. However, the short gap between EBRT and HDR-ILRT (3-14 days) and the short interfraction interval (three to seven days), may possibly explain the high incidence of the severe morbidity. In an optimization study on 35 patients, Akagi et al[14] treated patients with EBRT 50-61Gy followed by HDR-ILRT in 2 fractions of 4 or 5Gy (n=10) or 4-5 fractions of 2-2.5Gy (n=25) using a 15 mm or 20 mm diameter balloon applicator prescribed at 5 mm from the surface. They recommended the use of 4-5 fractions of 2-2.5Gy each for HDR-ILRT after 50-61Gy EBRT based on a significantly lower probability of late morbidity. In a report of 134 patients treated in 10 Italian centers with HDR-ILRT with or without EBRT, complication rates correlated with HDR-ILRT fraction size (9.5% complications for fraction sizes < 5 Gy, 20% with doses ranging 5-8 Gy and 38% with fraction doses > 8 Gy). Moreover, the esophagus was more severely injured when narrower tubes were used (24% with tube diameter < 2 mm, 19% with tubes diameter 2-6 mm and 5% with tube diameter > 6 mm).[15] In the present audit also, the use of narrower applicator i.e., 6 mm vs. the 10 mm diameter one was associated with a higher incidence of strictures although the mean duration of EBRT and HDR-ILRT interval was similar for those who did and did not experience development of late morbidity [Table - 3].

The use of concurrent chemotherapy with EBRT and HDR-ILRT has been reported in several studies. Calais et al[16] reported on 53 patients who received EBRT 60Gy/30fx and one to two weeks later 2 x 5Gy HDR-ILRT normalized at 5mm from the surface of a 10-14 mm diameter applicator. Concurrent chemotherapy consisted of CDDP 20 mg/m ² /day for four days, 5-FU 600 mg/m ² /day, continuous infusion for four days and mitomycin-C 6 mg/m ² on day one, planned for a total of three cycles beginning week one, four and -seven. Three cycles of chemotherapy could be completed by 64% of patients with 17% - grade III and 10% - grade IV hematological toxicity and a 2% fatality rate was observed. Strictures were seen in 13.5% and the authors stated that the low morbidity was due to total HDR-ILRT doses within 10-12 Gy, the use of a minimum applicator diameter of 10 mm and the dose prescription at 5 mm from the surface. In a retrospective study from Japan reported by Yorozu et al [17] patients were treated with a median EBRT dose of 50Gy followed -one to two weeks later by HDR-ILRT of 8-24 Gy in 2-4 fractions over two weeks prescribed at 5 mm from surface of a 15 mm diameter balloon applicator. Fifty-three patients also received CDDP 60 mg/m ² day 1 and 5-FU 600 mg/m ² continuous infusion day one to four given at the beginning and end of EBRT. Their outcome was compared with 116 historical controls without chemotherapy. Seventy percent complied with two cycles of chemotherapy and grade III and IV hematological toxicity seen in 32%. Late morbidity in the form of ulcers or strictures was seen in 12% of the historical group and in 34% of those receiving concurrent CT and RT ( P =0.01). Late morbidity in the CT+RT arm was clearly dependent on the HDR-ILRT doses: 0% for 8-12Gy, 11% for 16Gy and 25% for 20-24Gy. The RTOG reported on a phase I/II study of 50 patients who received concomitant CT+ EBRT and HDR-ILRT (RTOG 92-07), with the primary aim of intensifying radiation dose to reduce local failures.[7] Chemotherapy with 5-FU 1000 mg/m ² per 24 hour, 96-hour infusion along with CDDP 75 mg/m ² on day one was given in weeks one, five, eight and 11. EBRT (50Gy/25fx/5weeks) was started day one of CT and followed with HDR-ILRT 5Gy prescribed at 1 cm from center of a 4-6 mm diameter applicator on weeks 8, 9 and 10 for 80% of the patients. The first fraction of HDR-ILRT was given after 5-FU had been infused for 24-48h in week 8. The remaining 20% of the patients received two fractions of HDR-ILRT. Severe, life threatening toxicity and treatment related mortality occurred in 58%, 26% and 8% respectively, with 58% patients complying with the stipulated protocol. Six patients (12%) developed fistulas believed to be treatment related and half of these were fatal. The authors urged caution in employing ILRT as a boost following concurrent CT + EBRT. In the present audit, the use of concurrent CDDP as well as the combined use of EBRT+HDR-ILRT+CT resulted in a significantly higher incidence of strictures [Table - 3]. However, group E+I+C also had the highest proportion of patients treated with the 6 mm (narrowest) applicator [Table - 3]. Further it was seen that patients with better performance status had increased incidence of strictures, as perhaps they lived long enough to develop this late morbidity and this being a retrospective audit, selection bias is the most likely explanation [Table - 3]. In the multivariate model, the combined use of EBRT+HDR-ILRT+CT (as in group E+I+C) was retained as the only factor likely to influence both the development of ulcers (hazard ratio 10.36) and strictures (hazard ratio 4.2) [Table - 4].

In summary, when incorporating HDR-ILRT to CRT in treating cancer of the esophagus, the forgoing review of literature and results from the present audit reiterate the treatment related recommendations of the American Brachytherapy Society[18] These state that the interval between EBRT and ILRT should be at least two weeks when using concurrent chemotherapy; a small ILRT fraction size (as far as feasible) should be used; while a lower total dose of ILRT (within 10-12 Gy HDR) and an applicator with a diameter of 1cm result in the least morbidity. Therefore, this audit emphasizes the need for further optimization studies of brachytherapy before incorporating it in routine clinical practice so that the improved survival and sustained dysphagia relief that usually follows CRT protocols is not truncated by treatment induced stricture and ulcer development.

References

Copyright 2007 - Journal of Cancer Research and Therapeutics

The following images related to this document are available:

Photo images