Low-Dose Radiation Yields Lower Rates of Pathologic Response in Esophageal Cancer Patients

Background Although neoadjuvant chemoradiation (nCRT) followed by surgery is standard treatment for locally advanced esophageal or gastroesophageal junction (E/GEJ) cancer, the optimal radiation dose is still under debate. Objective The aim of this study was to assess the impact of different preoperative radiation doses (41.4 Gy, 45 Gy or 50.4 Gy) on pathologic response and survival in E/GEJ cancer patients. Methods All consecutive patients with E/GEJ tumors, treated with curative intent between January 2009 and December 2016 in two referral centers were divided into three groups (41.4 Gy, 45 Gy and 50.4 Gy) according to the dose of preoperative radiotherapy. Pathologic complete response (pCR) rates, postoperative morbidity, overall survival (OS) and disease-free survival (DFS) were compared among the three groups, with separate analyses for adenocarcinoma (AC) and squamous cell carcinoma (SCC). Results From the 326 patients analyzed, 48 were included in the 41.4 Gy group (14.7%), 171 in the 45 Gy group (52.5%) and 107 in the 50.4 Gy group (32.8%). Postoperative complication rates were comparable (p = 0.399). A pCR was observed in 15%, 30%, and 34% of patients in the 41.4 Gy, 45 Gy and 50.4 Gy groups, respectively (p = 0.047). A 50.4 Gy dose was independently associated with pCR (odds ratio 2.78, 95% confidence interval 1.10–7.99) in multivariate analysis. Within AC patients, pCR was observed in 6.2% of patients in the 41.4 Gy group, 29.2% of patients in the 45 Gy group, and 22.7% of patients in the 50.4 Gy group (p = 0.035). No OS or DFS differences were observed. Conclusions A pCR was less common after a preoperative radiation dose of 41.4 Gy in AC patients. Radiation dose had no impact on postoperative morbidity, long-term survival, and recurrence. Supplementary Information The online version contains supplementary material available at 10.1245/s10434-023-14810-8.

carboplatin-paclitaxel-based chemotherapy has become the standard of care after publication of the CROSS trial, offering pCR in 29% of all patients, 49% in squamous cell carcinoma (SCC) patients and 23% in AC patients. 6However, several nCRT combinations have been previously used, with variable results.The Swedish NeoRes trial reported a pCR rate of 28% with 40 Gy and platin-5-fluorouracil chemotherapy, 4 whereas Stahl et al. found a pCR rate of 15.6% with a 30 Gy RT dose. 5The Swiss SAKK 75/02 trial reported a pCR rate of 23% (38% for SCC and 16% for AC) after nCRT with docetaxel, cisplatin and 45 Gy, 7 whereas another retrospective series showed no differences in pCR rate comparing 41.4 Gy versus 50. 4 Gy for both SCC and AC. 8 Thus, despite the proven efficacy of different doses of preoperative radiation, comparative studies are scarce and results remain contradictory.This question becomes of particular clinical importance as the watch-and-wait strategy is gaining interest as a curative treatment option in esophageal cancer. 9van der Wilk et al. demonstrated similar rates of locoregional and distant recurrence for patients undergoing definitive chemoradiation and surgery-on-demand, compared with nCRT and upfront surgery. 10wo previous studies 11,12 assessed survival according to the preoperative RT dose (41.4 Gy, 45 Gy, or 50.4 Gy) but no significant differences were found.Although such small differences of RT doses are hardly be expected to entail a clear survival benefit, histological response represents a highly relevant outcome per se as it might increase the chances of a successful watch-and-wait strategy, especially for the more radio-resistant AC.The potential effect of increased radiation dose on postoperative and long-term morbidity needs to be considered as a higher dose of radiation may increase the risk of anastomotic leakage and surgical site infections. 13Late toxicities, such as radiationinduced pneumonitis, cardiotoxicity, and esophageal fibrosis have been reported in up to 20% of SCC patients with a 60 Gy/50.4Gy RT dose, 14 but this aspect remains poorly documented in patients with nCRT.
The aim of this study was to assess the impact of preoperative radiotherapy dose (41.4 Gy, 45 Gy, or 50.4 Gy) on pCR, postoperative morbidity, and long-term survival in patients with esophageal or gastroesophageal junction (E/ GEJ) tumors, with a subgroup analysis according to histological type.

PATIENTS AND METHODS
All consecutive patients with esophageal cancer treated with curative intent between January 2009 and December 2016 in two tertiary referral centers were retrospectively assessed.Inclusion criteria were AC and squamous cell cancer histological type, treated with neoadjuvant chemoradiotherapy (nCRT) and surgery.Patients with other histological types, emergency surgery, perioperative chemotherapy, salvage esophagectomy after definitive chemoradiotherapy, delayed surgery >15 weeks after the end of chemoradiation, and total radiation dose <41 Gy or >50.4 Gy were excluded from the study.Demographic, surgical, and oncological data were retrieved from prospectively maintained institutional databases.The primary outcome was defined as the histological response to treatment, whereas secondary outcomes were postoperative morbidity, overall survival, and DFS.The study was approved by the Ethics Committee of both participating centers (CER-VD ID v42017-02-17).

Tumor Staging and Treatment Details
Initial diagnostic work-up was performed with oesogastroduodenoscopy and biopsies, endoscopic ultrasound, thoracoabdominal CT scan, and total 18F-fludeoxyglucosepositron emission tomography/computed tomography (FDG-PET/CT) scan for detection of distant metastases.The 7th UICC/TNM system 15 was used for clinical staging.All cases were discussed in each institution's multidisciplinary tumor board to define the treatment.Local advanced lesions (cT3 and/or N+) underwent neoadjuvant treatment followed by surgery. 2 Restaging was performed 4 weeks after the end of treatment, with thoracoabdominal CT, FDG-PET/CT and endoscopy, and surgery was planned 4-8 weeks later.Specific assessment protocols for clinical complete response were not established as the watch-and-wait strategy was not a standard treatment option.Histologic response to treatment was assessed using the Mandard tumor regression grade (TRG) score, with TRG 1 representing pCR. 16Postoperative complications were graded according to the Clavien-Dindo system. 17hemotherapy protocols were based on 5-fluorouracilcisplatin and carboplatin-paclitaxel, while radiation dose ranged between 41.4 Gy and 50.4 Gy.Low-dose (41.4 Gy) radiation was introduced in the neoadjuvant context after publication of the CROSS trial in 2012. 6Radiation was administered to a total dose of 41.4 Gy, 45 Gy, or 50.4 Gy in fractions of 1.8 Gy.Although both participating centers shifted towards lower dose RT (41.4 Gy and 45 Gy) after 2012, universal treatment protocols were not imposed.The choice to maintain higher radiation doses in some patients was physician-dependent and not driven by a formal change of practice away from the CROSS regimen.In addition, several patients referred for surgery in the two participating centers had received nCRT in other institutions, where the CROSS regimen had not been clearly established.
At the time of this study, oncologic esophagectomy was performed by either the open or hybrid thoracoabdominal approach, as thoracoscopy was not yet introduced in current practice.The standard surgical approach was Low-Dose Radiation Yields Lower Rates … thoracoabdominal Lewis resection for middle-distal third tumors, and three-field McKeown resection for upper and middle-third lesions.

Statistical Analysis
Data were summarized using frequencies (%) for categorical variables, and median (interquartile range [IQR]) or mean (standard deviation [SD]) for continuous variables.Survival and recurrence were expressed as the median, in postoperative months (95% confidence intervals [CIs]).Intergroup comparisons were performed using the Chi-square or Fisher's exact test for categorical variables and analysis of variance (ANOVA) tests for continuous variables.A multivariate logistic regression model was used to define predictors of pathological complete response (pCR).Survival was assessed using the Kaplan-Meier method and the log-rank test, whereas a Cox regression model was used to identify variables independently related to overall survival (OS).Covariates with a p-value <0.1 on a univariate level were included in the multivariate analysis.Based on the different radio-sensibility of AC and SCC, subgroup analyses were performed according to histological type.Statistical analysis was performed using R studio version 1.1.383(Boston, MA, USA) and SPSS version 23.0 (IBM Corporation, Armonk, NY, USA) software.

Surgery and Postoperative Outcomes
A thoracoabdominal Ivor-Lewis resection was most commonly performed in all patients (p = 0.095).The hybrid (laparoscopic) approach was more frequently employed in the 41. 4 Gy group (77% vs. 50% in the 45 Gy group and 57% in the 50.4Gy group; p = 0.004).No differences were observed in terms of operative time, intraoperative blood loss or postoperative complications, particularly anastomotic leak rates and respiratory and cardiac complications.In-hospital mortality was also similar between groups (2%, 6%, and 3%, respectively; p = 0.352) [
Logistic regression was performed to identify independent predictors of pCR.Patients with 50.4 Gy had an increased chance of obtaining pCR compared with the 41.4 Gy group (adjusted odds ratio [OR] 2.78, 95% CI 1.10-7.99);among the other parameters, only cT status and cN status remained independent predictors of pCR (Table 4).

Overall Survival and Disease-Free Survival
OS did not show significant differences among the three groups, with a median of 30.0 months (95% CI 21-not available) for patients in the 41.4 Gy group, 31.6 months (95% CI 25.5-37.5)for patients in the 45 Gy group, and 30.0 months (95% CI 9.9-50.1)for patients in the 50. 4 Gy group (Fig. 2).In multivariate Cox regression, radiotherapy dose did not

DISCUSSION
In the present study, pCR was more common after highdose radiation (50.4 Gy or 45 Gy), especially in patients with AC, although this had no significant impact on longterm survival and recurrence rates.Dose escalation from 41.4 Gy to 50.4 Gy did not increase postoperative mortality or morbidity rates.Although 50% of esophageal cancer patients present with a curable disease stage, overall reported survival remains rather poor (20% 5-year OS). 18,19Since the early 2000s, neoadjuvant treatment (radiochemotherapy [nCRT] or chemotherapy [nCT]) followed by esophagectomy is the standard of care for locally advanced E/GEJ cancer in most Western countries. 20Early studies suggested that nCRT offered a survival advantage over nCT or surgery alone, however systemic treatment has known substantial progress in the meantime. 21,22Preliminary results of the randomized NeoAegis trial comparing nCRT versus nCT showed better histologic response in nCRT patients, without significant survival differences. 235][26][27] If histologic response represents a proxy of treatment efficacy and long-term outcomes, the adjunct of radiation in preoperative chemotherapy seems a reasonable approach.In this context, a better understanding of the optimal radiation dose in an nCRT context is needed.North American protocols recommend an nCRT dose of 50.4 Gy, 28,29 while European (European Society for Medical Oncology [ESMO]) guidelines suggest nCRT with 41.4 Gy for SCC, but do not clearly specify radiation dose in AC, although the low-dose CROSS protocol is also suggested. 2The present study illustrates suboptimal histological response rates with low-dose radiation in AC patients.Previously, Nabavizadeh et al. showed similar pCR after the modified CROSS regimen, using 50.4Gy compared with the standard low dose, despite an increased risk of severe radiation-induced acute lung injury in the higher dose. 30In another study including 80% AC and 20% SCC, Ji et al showed that low-dose radiation (41.4 Gy) may even offer improved OS compared with higher-dose regimens, with similar local control and cT and N status downstaging. 31Although the reason behind the survival benefit observed in the low-dose group remains unclear, increased treatment-related mortality may be part of the explanation.
Treatment-related toxicity is an important issue to consider when defining the optimal radiation dose.In the present study, postoperative morbidity/mortality rates were not increased in the high-dose radiation group, however specific data on nCRT toxicity are not available.It has previously been reported that up to 14% of patients who started nCRT for esophageal cancer will not be able to proceed to surgery due to disease progression or reduction of physical functioning and treatment-related toxicity. 32In a definitive chemoradiation context, the French FREGAT group had suggested that when radiation dose escalates to 55 Gy, patients experienced increased major postoperative morbidity, anastomotic leakage rates and mortality, and even reduced OS. 33Stahl et al. showed equivalent survival in SCC patients treated with exclusive high-dose chemoradiation (65 Gy) alone versus low-dose nCRT followed by surgery (40 Gy), with less treatment-related mortality in the nCRT arm. 34he CONCORDE/PRODIGE 26 trial showed that 66 Gy, although not entailing more toxicity than 50 Gy, failed to improve progression-free survival, proposing 50 Gy as a standard dose for definitive CRT. 35Thus, dose escalation to >50 Gy is nowadays discouraged, whereas more recent techniques, such as the Proton Beam Therapy (PBT) have emerged to reduce off-target adverse effects by focusing dose distribution to the primary tumor. 36he current study has particular clinical relevance in the era of the 'watch-and-wait' treatment for esophageal cancer, where obtaining the best possible response after neoadjuvant treatment is key.Indeed, in patients with complete clinical response (cCR) after nCRT, delayed surgery on-demand has been suggested in case of local relapse, as it provided similar survival results to nCRT and upfront surgery. 10The 'watchand-wait' strategy seems very promising, 9 especially in the SCC histology, as the CROSS trial showed a pCR in almost 50% of SCC compared with 25% of AC patients. 6Although the challenge to reliably detect cCR after chemoradiation needs to be acknowledged, 37 the SANO trial is currently assessing the watch-and-wait strategy in patients presenting cCR after nCRT with 41.4 Gy. 38,39 Based on the present study's results, the CROSS protocol radiation dose (41.4 Gy) might carry a high risk of undertreating AC lesions, as significantly lower rates of pCR were observed compared with 45-50.4 Gy.
This study has some limitations that need to be discussed.Although data are extracted from prospectively maintained databases, the retrospective character of the study entails the shortcoming of missing data, especially in terms of treatmentrelated toxicity.In addition, some baseline differences exist among the three radiotherapy groups, notably in the histological type, tumor location, cT stage, and chemotherapy regimen.To face this methodological drawback, we performed separate subgroup analyses by histological type for all the main outcomes (pCR, survival) due to the high clinical significance of histology.In addition, rigorous multivariable analyses adjusted for all the above-mentioned confounders when assessing the independent predictive value of radiotherapy dose on pCR and survival.Another limitation concerns the absence of clearly defined criteria for radiotherapy regimen choice.The chosen study period is anterior to recent advances in systemic treatment of esophageal cancer, such as adjuvant checkpoint inhibitors, that led to improved DFS after nCRT.Finally, radiotherapy modalities are becoming increasingly precise and efficient, and the 50.4Gy dose previously used in the neoadjuvant setting is now outdated and only admitted as definitive CRT.However, as our data illustrate, this dose might still have its place in selected cases, as, for example, an AC patient in watch-andwait treatment strategy.

CONCLUSION
Our findings suggest that low-dose radiation is efficient in SCC; a higher dose of 45-50.4Gy may be needed for AC as it offers increased chances for pathological complete response, without compromising postoperative outcomes.These findings are of particular clinical relevance in candidates for the watch-and-wait strategy as they suggest low-dose RT to be potentially insufficient for AC lesions.

TABLE 1
AC adenocarcinoma, ASA American Society of Anesthesiologists, GEJ gastroesophageal junction, SCC squamous cell cancer, SD standard deviation a Percentage of weight loss on diagnosis compared with baseline weight

TABLE 2
Surgical characteristics and postoperative outcomes according to preoperative radiation dose Data are expressed as frequency (%) or mean [standard deviation] as appropriate a Defined as Clavien-Dindo grade >IIIa