Introduction

Esophageal cancer (EC) is the seventh most common cancer and ranks sixth in terms of cancer-associated mortality worldwide,1 with developing countries accounting for approximately 80% of cases and deaths.2 The mainstay of treatment for locally advanced EC is surgery, but the survival outcomes following surgery alone are dismal, with 5-year survival rates of approximately 25%.3 Following the “ChemoRadiotherapy for Oesophageal cancer followed by Surgery Study” (CROSS)4 and the “NEOCRTEC5010”5 trials, neoadjuvant chemoradiotherapy (NACRT) and surgery have become the standard trimodality therapy for locally advanced but resectable EC. However, a significant number of patients suffer from cancer-associated malnutrition and cachexia, which could lead to deterioration of their general well-being during the preoperative therapy, resulting in treatment interruption/poor tolerance and poor candidature for surgery. Additionally, there may be a general reluctance to practice this form of preoperative treatment due to the fear of increased postoperative complications and the lack of adequate infrastructure for its delivery, particularly in low-middle-income nations.6 Furthermore, few reported literature from India and other South Asian countries often shows poor patient tolerability to NACRT, particularly to its chemotherapy component.78 The advanced stage at presentation, poor baseline performance and nutritional statuses of the patients, and predominance of esophageal squamous cell carcinoma (ESCC) are common in Asian countries291011; these factors may alter the behavior and response of the tumor to NACRT and could affect its effective delivery.

NACRT, as per the CROSS regimen, is the most accepted preoperative therapy for ESCC. However, real-world data from low-middle-income nations about the selection and outcomes of ESCC patients receiving NACRT as per the CROSS regimen in the standard care setting outside the scope of a clinical trial are scarce. Although there are a few studies from the Indian subcontinent that report the feasibility, safety, and outcomes of NACRT in operable EC,781012 data are scarce on the pretherapy factors that could influence the treatment tolerability8 and the effect of the completion of NACRT on the pathological outcomes. Identifying these pretherapy factors could affect the choice of trimodality therapy or serve as a guide for its modifications. This can also help in better optimization of patients receiving NACRT and improve their treatment tolerance.

The purpose of this single-center study was to explore the feasibility, safety, and tolerability of NACRT (CROSS protocol [paclitaxel–carboplatin + 41.4 Gy] or its modification [45 Gy]) in patients with ESCC and to determine the pretherapy factors that could potentially affect the treatment completion. Additionally, we studied the impact of neoadjuvant treatment completion on the various pathological outcomes.

Patients and Methods

After obtaining the ethics committee approval (Min. No. 12998/2020), a retrospective review of patients with ESCC, who were discussed in the tumor board of our cancer center from January 2015 to May 2022, was performed. The relevant data were collected from a prospective database and the electronic medical records.

The clinico-demographic details collected included details of comorbidities, substance abuse, body mass index, baseline laboratory data, tumor location, staging, and treatment received. The comorbidities were graded using the Charlson comorbidity index (CCI),13 and a CCI >2 was considered a “high” score. The patient's performance status was assessed using the Eastern Cooperative Oncology Group (ECOG) score.14 The diagnosis of EC was confirmed using esophagogastroscopy and biopsy. A contrast-enhanced computed tomography of the thorax/abdomen was used to stage the disease. ¹⁸Fluorodeoxyglucose positron emission tomography was performed selectively. The Union for International Cancer Control TNM (tumor node metastasis) classification (8th edition) was used for disease staging.15

The initial part of this study cohort includes patients who received 45-Gy radiation and concurrent paclitaxel/carboplatin × 5 cycles (defined as “modified” CROSS regimen [mCROSS]), which was followed in our center before the adaptation of the CROSS protocol. The remaining patients received NACRT as per the CROSS protocol (41.4 Gy in 23 fractions and concurrent 5 cycles of paclitaxel [50 mg/m²] and carboplatin [area under the curve (AUC) = 2]).4 Total radiation of 41.4 to 45 Gy is widely accepted as “low-dose radiation” in a neoadjuvant setting for EC; hence, we decided to include the former group in this study. Patients who received NACRT beyond the original CROSS trial inclusion criteria4 were classified as an “extended” criteria group.

Radiation was planned to include the gross tumor and the nodal disease with appropriate margins. The adverse events were reported using Common Terminology Criteria for Adverse Events (CTCAE) v4.0.16 Chemotherapy completion was defined as having received a minimum of four cycles.

The clinical reassessment was performed about 6 weeks after NACRT, and suitable patients underwent surgery. The reasons to defer the operation were documented, if any. The severity of 30-day postoperative complications was graded using the Clavien–Dindo grading.17 The pathological (ypTNM) staging was defined as per UICC TNM (8th edition).15 The completeness of the oncological resection and the tumor regression grading (TRG) was defined as per the College of American Pathologists criteria.15

Continuous variables were documented using mean with standard deviation (SD) or median with range or interquartile range (IQR). The frequencies and percentages were used to document the categorical variables, which were compared using an independent t-test or Mann–Whitney U test. The chi-squared test or Fisher exact test was used to find the association between completion of chemotherapy (≥4 vs. <4 cycles) and the pretherapy factors that could influence the treatment completion. The study variables that were significant at less than 0.05 levels in an unadjusted analysis were included in a multivariable logistic regression model. The model fit was assessed using the Hosmer and Lemeshow goodness fit test. A p-value less than 0.05 was considered significant. All statistical analyses were performed using Statistical Package for the Social Sciences (SPSS), version 21.0 (IBM Corp., Armonk, NY, United States).

Results

Discussion

In the CROSS trial,4 which had randomized patients with operable EC to receive either NACRT followed by surgery or surgery alone, the treatment adherence was excellent, with chemotherapy and radiotherapy being tolerated by 91 and 92% of patients, respectively. Although patients in this trial tolerated the paclitaxel–carboplatin regimen well, adapting this regimen in a real-life scenario needs caution, as not every patient of operable EC is medically fit to receive the combined modality treatment. Toxopeus et al18 compared the outcomes of patients receiving the CROSS protocol NACRT outside the trial with those within the CROSS trial. There were older patients with frequent comorbidities, poor performance status, and advanced “N” stage in the post-CROSS cohort. Despite these differences, greater than 95% of patients in each group tolerated the entire chemotherapy course, and there was no significant difference in adverse events or survival between the groups. Another study by de Heer et al19 assessed the impact of the “extended” CROSS eligibility criteria in patients with EC. Although there was no effect on the NACRT-associated toxicity and postoperative safety, the prognosis was adversely affected by the application of the “extended” criteria. In another study by Paireder et al,20 among the 56 patients receiving the mCROSS protocol (46 Gy), 64% of patients progressed to surgery, and 83% of them could complete ≥4 chemotherapy cycles, with grade 3 leukopenia (34%) being the foremost reason for noncompletion of chemotherapy.

It must be noted that the dominant histology of EC in the CROSS trial and the above post-trial cohort studies was adenocarcinoma (>70% of patients). In contrast, squamous carcinoma is the dominant histology in Asian countries.2910 Patients with poor performance, comorbidities, significant weight loss, and locally advanced, long-segment diseases are frequent in low-middle-income countries.267 In the NEOCRTEC5010 trial,5 comparing upfront surgery versus NACRT and surgery for ESCC, 40 Gy of radiotherapy and two cycles of concurrent vinorelbine and cisplatin were given to patients in the NACRT arm; 87% completed the chemotherapy, and all except two patients received the planned radiotherapy. Although the study had significant (grade 3/4) hematological complications (neutropenia [45.7%], leucopenia [48.9%]), chemotherapy was withheld only if the complication persisted for 2 weeks. Chemotherapy cycles being fewer and 3 weeks apart are the probable reasons that patients were able to recover from cytopenia and complete their chemotherapy cycles.

The published literature on real-life experience with NACRT for ESCC is limited, particularly concerning the factors affecting its completion. Wong et al11 observed that among ESCC patients receiving the CROSS protocol chemoradiotherapy, 91% of patients could complete ≥4 cycles of chemotherapy. Krishnamurthy et al7 and Bajwa et al10 analyzed the outcomes of patients from the Indian subcontinent with EC (SCC dominant cohorts) receiving NACRT as per the CROSS regimen; ≥4 chemotherapy cycles were completed by 80 and 97% of patients, respectively. However, the authors did not report the proportion of patients who were “eligible” or “extended” as per the CROSS trial inclusion criteria. Strict adherence to patient inclusion criteria in these studies likely led to the high chemotherapy tolerance seen in their experience. It should also be noted that these studies included patients with adenocarcinoma, whereas, in our study, only patients with ESCC were included.

This study aimed to analyze the real-world experience concerning the feasibility and tolerability of NACRT for the CROSS/mCROSS regimen in patients with ESCC. It also evaluated the factors predicting the completion of NACRT. In our experience, radiotherapy was well tolerated without treatment breaks; however, the tolerability of chemotherapy was poorer, with neutropenia being the commonest reason for the omission of chemotherapy in 61% of patients. The application of patient inclusion criteria for the CROSS trial in our cohort showed that nearly 45% of patients would have been “ineligible” for receiving NACRT as per the original trial protocol. We had to include these patients since many presented with locally advanced/long-segment disease in our population. Although there was no statistically significant difference in the tolerance to chemotherapy between the CROSS “eligible” and “extended” criteria groups (51.7 vs. 55.1%, p = 0.721), it has to be noted that nearly 50% of patients even within the CROSS “eligible” group failed to complete at least four cycles. Although there were no significant differences in the baseline demographic factors, compared with the CROSS trial, our study had patients with ESCC alone, and the tumors were relatively more advanced with longer tumor length (mean tumor length: 5.5 cm), necessitating a larger radiotherapy target volume. These factors perhaps contributed to a high risk of bone marrow toxicity/myelosuppression within the CROSS “eligible” group, thereby leading to a higher incidence of neutropenia.

Leukopenia and thrombocytopenia are frequent adverse effects encountered during NACRT.21 Drug dose modifications, chemotherapy interruptions, and early treatment discontinuation are frequent in patients developing neutropenia, and this emerged as the leading cause of chemotherapy discontinuation in this study. In a few of our patients, the occurrence of neutropenia after the fourth chemotherapy pushed us to avoid the last cycle of chemotherapy, although the patient might have recovered from neutropenia. Such a decision was taken in patients in whom there were frequent interruptions in the treatment or severe neutropenia after the fourth cycle, which could potentially delay the surgery, which is considered the key cancer treatment. Although PEGylated recombinant human granulocyte colony-stimulating factor can be considered judiciously in patients developing neutropenia,22 its routine use to prevent or treat chemotherapy-associated myelosuppression is not advocated in patients with EC receiving NACRT. We do not use this agent in our routine practice.

Radiation-induced esophagitis often interrupts treatment in patients with EC.821 Although radiation-induced esophagitis did occur in our patients, it led to treatment interruption only in 2% of our patients. Patients with dysphagia to solids often have a nasogastric tube placed for feeding prior to NACRT to ensure nutritional maintenance during therapy and to manage radiation esophagitis–related dysphagia. In this study, 34% of patients alone had a nasogastric tube placed before or during NACRT. Despite our advice for tube feeding in all patients with significant dysphagia and/or weight loss, patients' acceptability was poor due to their general reluctance to have a nasal tube for a prolonged duration and due to local social taboos.

As the tolerability of chemotherapy was poor in this study, we studied the pretherapy factors that could influence the completion of chemotherapy. Multivariate analysis showed that patients with anemia had a lower chance of completing chemotherapy. Although the extent of anemia leading to treatment interruption needs future validation, anemia needs timely correction before and during NACRT. In our study, nearly a third of patients had at least one medical comorbidity, and a high CCI (>2) predicted chemotherapy discontinuation. These findings demonstrate that CCI may be a potential indicator for pretreatment patient stratification and periodic assessment, and timely optimization of the patient's medical diseases and functional status can improve treatment tolerance. Additionally, those with a high CCI can be offered alternate treatment options, for instance, radical chemoradiotherapy for ESCC. However, the findings from our study concerning the role of CCI need validation from future large-sized studies. Conventional radiotherapy delivery techniques, particularly when combined with a relatively more toxic chemotherapy regimen, can lead to poor patient tolerance to NACRT.8 The newer techniques of radiation delivery, mainly intensity-modulated radiation therapy (IMRT), can minimize collateral tissue damage, including bone marrow and cardiopulmonary toxicity.23 In our study, the application of image-guided conformal techniques resulted in superior tolerability of concurrent chemotherapy compared with conventional techniques. However, despite our study's frequent use of the conformal IMRT technique, chemotherapy tolerance was still poor compared with the CROSS trial, which utilized the three-dimensional conformal radiation therapy (3D-CRT) technique alone.4 However, in a study by Innocente et al,24 utilizing radiation doses of 52.5 to 54 Gy (IMRT technique) along with five cycles of concurrent paclitaxel and carboplatin to patients with EC (SCC, 58%), 54% of patients could complete at least four cycles of chemotherapy, despite receiving a higher intensified dose than reported in our study. This study's findings demonstrate that the treatment toxicity and, hence, the chemotherapy component's tolerability could be improved by better tumor targeting using newer radiation delivery techniques. Although shown to be beneficial in improving the tolerability of NACRT24 and reducing postoperative morbidity,23 the lack of wide availability and the increased cost of advanced radiation techniques are significant limiting factors for their routine application in low-middle-income countries.

The real-life incidence of planned surgery cancellations after neoadjuvant therapies for EC can be 16 to 22%.25 In our experience, nearly 25% of our patients did not receive esophagectomy, which is higher than the figure of 6% reported from the CROSS trial4 but better than the figure of 57% reported by Anap et al12 among patients with EC. In the study by Krishnamurthy et al, 78% of patients received operation and 14% were deemed inoperable on-table.7 In the NEOCRTEC5010 trial, the most common reason for deferring surgery was patient refusal (12.9% of patients), whereas disease progression or poor performance precluded surgery in 0.8% of patients alone.5 Unresectability, patient refusal, mainly due to complete clinical response, and poor general condition are reported to be frequent reasons for not performing surgery in patients with ESCC treated with NACRT.25 In our experience, among those who did not reach surgery, patient refusal (30%), disease progression/inoperability (30%), and lack of fitness due to comorbid illnesses or poor performance (26%) were the frequent reasons for deferring surgery. Low socioeconomic background2 and lack of socioeconomic support are common contributing factors for deferring surgery in patients from the Indian subcontinent.12 Hence, cancer care providers in a low-middle-income countries should also consider the patients' socioeconomic support before finalizing their decision regarding multimodality treatment, and the patients who decide to defer surgery due to clinical improvement must be thoroughly counseled about the pros and cons of such decisions.

In the neoadjuvant settings, concurrent chemotherapy acts primarily as a radiosensitizer. Through its radiosensitizing effect, it is shown to reduce the locoregional and distant metastasis in EC, compared with radiotherapy alone.26 Hence, completing the chemotherapy component is vital to improve EC's pathological and survival outcomes. In our experience, the rate of R0 resection, pCR, and node positivity rates were similar to that reported in the literature.451027 Although there was a trend toward better pathological outcomes in patients completing chemotherapy, these differences were statistically insignificant. The omission of a few cycles of concurrent chemotherapy did not affect various pathological outcomes, particularly the histological response; however, we feel that there is an actual positive effect, which we were unable to demonstrate, perhaps due to small individual samples, and hence this needs further prospective evaluation. Whether a survival advantage exists for the chemotherapy completion group (vs. the noncompletion group) also needs careful exploration through further long-term follow-up of our cohort and future prospective studies. Additionally, should we be looking at adjuvant therapy for patients in whom the preoperative chemotherapy cycles could not be completed, particularly if the pathological tumor response is suboptimal? This question gains relevance and importance in the current era of possible immunotherapy options.

Our study had a few limitations. First, it was a retrospective, single-center study with inherent biases. Second, the psychosocial issues limiting the completion of preoperative treatment are not studied, considering most patients were from low socioeconomic backgrounds. “Prehabilitation” programs, including dietary interventions, can help improve overall treatment compliance and reduce treatment-associated morbidity in esophagogastric cancer patients.28 However, data on baseline nutritional assessment and degree of weight loss were lacking for most of our patients; hence, the effect of pretherapy nutritional inadequacy and weight loss or its improvement after nutritional interventions on the treatment completion could not be studied in our cohort. This was the third limitation of this study. Fourth, since the proportion of patients who received five cycles of chemotherapy was low, completion of a minimum of four cycles was taken as the cutoff to study the oncological effects of chemotherapy completion. Furthermore, the superiority of various conformal radiation delivery techniques in terms of their tolerability and toxicity could not be established due to the small individual sample size. Finally, the long-term oncological effects of chemotherapy completion and the oncological outcomes of patients who never reached surgery need to be explored. Despite these limitations, our study is one of the largest studies from the subcontinent, which has explored the pretherapy factors contributing to chemoradiotherapy completion in a cohort of ESCC patients.

Conclusion

Tolerability to the CROSS regimen or its modification is poor among the Indian subcontinent patients with esophageal squamous cancer. Although radiotherapy is well tolerated, chemotherapy is often associated with hematological toxicity, which leads to suspension or omission of therapy. Anemia at presentation and a high CCI are strongly associated with noncompletion of the planned chemotherapy cycles, and these indicators can guide us in better case selection for NACRT or its modification. Conformal radiation delivery techniques can help reduce toxicity and improve treatment tolerance. Although the completion of chemotherapy has no significant effect on the various pathological outcomes, its impact on long-term survival in squamous carcinoma needs to be prospectively studied.