Association between Submucosal Fibrosis and Endoscopic Submucosal Dissection of Recurrent Esophageal Squamous Cell Cancers after Chemoradiotherapy

Simple Summary The efficacy and safety of endoscopic submucosal dissection for early esophageal cancer after chemoradiotherapy have not been established. In this study, we focused on the fibrosis of the submucosa. As a result, we found that endoscopic submucosal dissection for early esophageal cancer can be performed reliably without adverse events, but the procedure takes longer for lesions with strong fibrosis of the submucosa. Abstract Endoscopic resection is a treatment of choice for a metachronous early-stage esophageal squamous cell carcinoma (ESCC) appearing after a radical cure of esophageal cancer by chemoradiotherapy (CRT). However, non-curative resection, and procedural complications including perforation due to radiation-induced submucosal fibrosis, are a concern. This study aimed to evaluate the association between submucosal fibrosis and the usefulness and safety of endoscopic submucosal dissection (ESD) in ESCC after CRT. This study retrospectively analyzed 13 lesions in 11 patients in our institute. Submucosal fibrosis under the lesion (F score) was classified into three levels (F0: none or mild, F1: moderate, and F2: severe) based on endoscopic and histopathologic findings. All lesions were F1 or greater (F1: 8 lesions and F2: 5 lesions). En bloc and R0 resection rates were both 100%. The procedural speed was slower in F2 than in F1 (F1 vs. F2; 15.1 mm2/min vs. 7.1 mm2/min, p = 0.019), without procedure-related adverse events. At a median follow-up of 42 months (range: 14–117 months) after ESD, 7 of 11 (63.6%) patients were alive without recurrence, and without ESCC-related death. ESCC after CRT reliably and safely resected en bloc by ESD but was more difficult in lesions with strong submucosal fibrosis.


ESD Procedure and Pathological Evaluation
ESD was performed under the general anesthesiologist's control. All ESDs were performed by expert physicians who had performed > 50 esophageal ESDs, or by non-expert physicians under the supervision of expert physicians, according to previous reports [26][27][28][29][30][31][32].
An endoscope (GIF-Q260J or GIF-H290T; Olympus Medical Systems Corp., Tokyo, Japan) was used with carbon dioxide, and A VIO300D or VIO3 (ERBE Elektromedizin, Tübingen, Germany) was used as the high-frequency system. The area around the lesion was marked using a Dual knife (Olympus Medical Systems Corp.), which was then used to incise the mucosa after injecting 0.4% sodium hyaluronate (MucoUp; Boston Scientific Japan, Tokyo, Japan) into the submucosa using a needle (ImpactFlow; TOP Corp., Tokyo, Japan) [33][34][35]. An IT knife nano (Olympus Medical Systems Corp.) was mainly used for submucosal dissection, and SB knife Jr (SB Kawasumi, Co., Ltd., Tokyo, Japan), a scissorstype knife, was used in cases of severe submucosal fibrosis. Furthermore, coagulation was performed using hemostatic forceps (Coagrasper; Olympus Medical Systems Corp.) for bleeding and hemorrhage prevention.
Triamcinolone is injected into the mucosal defect immediately after resecting the lesion to prevent postoperative stenosis if the circumference of the mucosal defect after resection was more than 3/4 circumference [36][37][38].

Outcomes
The en bloc resection rate, R0 resection rate, procedural speed, procedure-related adverse events, and prognosis were evaluated. The procedural speed (mm 2 /min) was calculated by dividing the dissection area (mm 2 ) by the procedure time (min). The procedure time was defined as the time from the start of the mucosal incision to the end of lesion resection, and the dissection area was calculated as the radius of the long axis × radius of the short axis × 3.14.
Regarding procedure-related adverse events, perforation was defined as endoscopic confirmation of the mediastinum or free air on computed tomography (CT). CT was performed only when endoscopically suspected perforation was detected during ESD. Postoperative bleeding was defined as the presence of hematemesis or black stools after ESD and active bleeding or exposed blood vessels on endoscopy [40].
Submucosal fibrosis under the lesion was named as "F score." The F score based on endoscopic and pathological findings was defined as the endoscopic F (eF) score and pathological F (pF) score, respectively. The eF score was evaluated by two endoscopists (T.K. and R.K.) based on the stored endoscopic images and videos and was assigned to 0 (none or mild fibrosis), 1 (moderate fibrosis), and 2 (severe fibrosis) by agreement between the two endoscopists ( Figure 1) [41]. The pF score was evaluated by one pathologist (Y.O.) as 0 (none or mild fibrosis), 1 (moderate fibrosis), and 2 (severe fibrosis) based on the Elastica Masson staining of the ESD specimen ( Figure 2) [15]. Finally, the higher score between eF and pF scores was defined as the F score of the lesion ( Figure 3). assess submucosal fibrosis.

Outcomes
The en bloc resection rate, R0 resection rate, procedural speed, procedure-related adverse events, and prognosis were evaluated. The procedural speed (mm 2 /min) was calculated by dividing the dissection area (mm 2 ) by the procedure time (min). The procedure time was defined as the time from the start of the mucosal incision to the end of lesion resection, and the dissection area was calculated as the radius of the long axis × radius of the short axis × 3.14.
Regarding procedure-related adverse events, perforation was defined as endoscopic confirmation of the mediastinum or free air on computed tomography (CT). CT was performed only when endoscopically suspected perforation was detected during ESD. Postoperative bleeding was defined as the presence of hematemesis or black stools after ESD and active bleeding or exposed blood vessels on endoscopy [40].
Submucosal fibrosis under the lesion was named as "F score." The F score based on endoscopic and pathological findings was defined as the endoscopic F (eF) score and pathological F (pF) score, respectively. The eF score was evaluated by two endoscopists (T.K. and R.K.) based on the stored endoscopic images and videos and was assigned to 0 (none or mild fibrosis), 1 (moderate fibrosis), and 2 (severe fibrosis) by agreement between the two endoscopists ( Figure 1) [41]. The pF score was evaluated by one pathologist (Y.O.) as 0 (none or mild fibrosis), 1 (moderate fibrosis), and 2 (severe fibrosis) based on the Elastica Masson staining of the ESD specimen ( Figure 2) [15]. Finally, the higher score between eF and pF scores was defined as the F score of the lesion ( Figure 3).   The submucosa looks slightly white, but the submucosal elevation is possible. (c) Severe fibrosis (eF2). The submucosa looks highly white, and submucosal elevation is inadequate. assess submucosal fibrosis.

Outcomes
The en bloc resection rate, R0 resection rate, procedural speed, procedure-related adverse events, and prognosis were evaluated. The procedural speed (mm 2 /min) was calculated by dividing the dissection area (mm 2 ) by the procedure time (min). The procedure time was defined as the time from the start of the mucosal incision to the end of lesion resection, and the dissection area was calculated as the radius of the long axis × radius of the short axis × 3.14.
Regarding procedure-related adverse events, perforation was defined as endoscopic confirmation of the mediastinum or free air on computed tomography (CT). CT was performed only when endoscopically suspected perforation was detected during ESD. Postoperative bleeding was defined as the presence of hematemesis or black stools after ESD and active bleeding or exposed blood vessels on endoscopy [40].
Submucosal fibrosis under the lesion was named as "F score." The F score based on endoscopic and pathological findings was defined as the endoscopic F (eF) score and pathological F (pF) score, respectively. The eF score was evaluated by two endoscopists (T.K. and R.K.) based on the stored endoscopic images and videos and was assigned to 0 (none or mild fibrosis), 1 (moderate fibrosis), and 2 (severe fibrosis) by agreement between the two endoscopists ( Figure 1) [41]. The pF score was evaluated by one pathologist (Y.O.) as 0 (none or mild fibrosis), 1 (moderate fibrosis), and 2 (severe fibrosis) based on the Elastica Masson staining of the ESD specimen ( Figure 2) [15]. Finally, the higher score between eF and pF scores was defined as the F score of the lesion ( Figure 3).   Evaluation criteria for pathological F score. (a) None or mild fibrosis (pF0). Fibrosis was defined as mild when fibrous tissue was <20%. (b) Moderate fibrosis (pF1). Fibrosis was defined as moderate when fibrous tissue was between 20% and 60%. (c) Severe fibrosis (pF2). Fibrosis was defined as severe when fibrous tissue was 60% or more.
Regarding the patient characteristics, the disease stage before CRT in patients undergoing definitive CRT (dCRT) was evaluated based on esophagogastroduodenoscopy (EGD), endoscopic ultrasound, CT, and positron emission tomography (PET). Patients who underwent CRT as an additional treatment after ESD were evaluated based on the histopathological results of ESD, CT, and PET. Patients were followed up with EGD and CT once or twice a year after ESD. moderate when fibrous tissue was between 20% and 60%. (c) Severe fibrosis (pF2). Fibrosis was defined as severe when fibrous tissue was 60% or more. Regarding the patient characteristics, the disease stage before CRT in patients undergoing definitive CRT (dCRT) was evaluated based on esophagogastroduodenoscopy (EGD), endoscopic ultrasound, CT, and positron emission tomography (PET). Patients who underwent CRT as an additional treatment after ESD were evaluated based on the histopathological results of ESD, CT, and PET. Patients were followed up with EGD and CT once or twice a year after ESD.

Statistical Analysis
All identified patients were analyzed, and subgroup analyses by submucosal fibrosis were conducted. Values are reported as medians with ranges. Statistically significant differences between patient characteristics and ESD results were assessed using the Mann-Whitney U test for continuous variables. Differences were significant at p-values of < 0.05. This analysis was performed using the Statistical Package for the Social Sciences software (version 21 for windows; IBM Corp., Armonk, NY, USA).

Patients and Lesion Characteristics
This study analyzed 13 lesions in 11 patients (Table 1). CRT was for dCRT in 8 cases and additional therapy after ESD in 3 cases, with a median total radiation dose of 60 Gy in a median follow-up of 38 months (range: 13-85 months) from the end of CRT to ESD.

Statistical Analysis
All identified patients were analyzed, and subgroup analyses by submucosal fibrosis were conducted. Values are reported as medians with ranges. Statistically significant differences between patient characteristics and ESD results were assessed using the Mann-Whitney U test for continuous variables. Differences were significant at p-values of <0.05. This analysis was performed using the Statistical Package for the Social Sciences software (version 21 for windows; IBM Corp., Armonk, NY, USA).

Patients and Lesion Characteristics
This study analyzed 13 lesions in 11 patients (Table 1). CRT was for dCRT in 8 cases and additional therapy after ESD in 3 cases, with a median total radiation dose of 60 Gy in a median follow-up of 38 months (range: 13-85 months) from the end of CRT to ESD. There were no cases of residual recurrence after CRT, and all ESCCs that underwent ESD were new lesions with metachronous recurrence. ESD: endoscopic submucosal dissection, CRT: chemoradiation therapy, F score: submucosal fibrosis score, eF: endoscopic fibrosis score, pF score: pathological fibrosis score, dCRT: definitive chemoradiotherapy, aCRT: additional therapy after ESD Ce: cervical esophagus, Ut: upper thoracic esophagus, Mt: middle thoracic esophagus, Lt: lower thoracic esophagus. SCC: squamous cell carcinoma, EP: epithelium, LPM: lamina propria mucosae, MM: muscularis mucosae. Case 1-1 and case 1-2 were the recurrent lesions identified 26 months after CRT in the same patient, both of which were treated on the same day. Case 2-1 and case 2-2 were recurrent lesions found in the same patient and treated 37 months and 49 months after CRT. * Radiation was discontinued at up to 40 Gy due to adverse events.

Treatment Outcomes of ESD
The cancer depth was 4 (30.8%) with the epithelium, 7 (53.8%) with the lesions on the lamina propria mucosae, and 2 (15.4%) with muscularis mucosae. HM, VM, and LVI were negative in all lesions. Both en bloc resection rate and R0 resection rate were 100%. Five lesions with mucosal defect circumference of more than 3/4 after ESD were treated with triamcinolone injection and no stenosis occurred. No other adverse events, such as perforation or postoperative bleeding occurred (Table 1).

Prognosis
At a median follow-up of 42 months (range: 14-117 months) after ESD, 7 of 11 patients (63.6%) were alive without recurrence. There were no ESCC-related deaths, and the causes of death in the four other cases were senility, head and neck cancer, liver cirrhosis, and debilitating death due to alcoholism (Figure 4).  Procedure-related adverse events, n 0 0 0 -* p-values were calculated using the Mann-Whitney U test. ESD: endoscopic submucosal dissection, CRT: chemoradiation therapy. Procedure speed (mm 2 /min); dissection area (radius of long axis × radius of short axis × 3.14 (mm 2 )/procedure time (min).

Prognosis
At a median follow-up of 42 months (range: 14-117 months) after ESD, 7 of 11 patients (63.6%) were alive without recurrence. There were no ESCC-related deaths, and the causes of death in the four other cases were senility, head and neck cancer, liver cirrhosis, and debilitating death due to alcoholism (Figure 4).

Discussion
This study revealed that ESD for ESCC that recurred metachronously in the irradiated area after CRT is safe and reliable for resection. Additionally, the evaluation of sub-

Discussion
This study revealed that ESD for ESCC that recurred metachronously in the irradiated area after CRT is safe and reliable for resection. Additionally, the evaluation of submucosal fibrosis under ESCC after CRT, based on endoscopic and histopathologic findings, is considered novel because the procedural speed was slow in lesions with significant fibrosis.
Surgery and photodynamic therapy (PDT) were performed for locally recurrent ESCC after CRT in esophageal cancer [42][43][44][45]. However, surgical procedures for ESCC after CRT are highly invasive with reported perioperative mortality rates as high as 7.4-25% [42,[46][47][48][49][50][51]. PDT is advantageous because it treats ESCC with strong fibrosis, but its downside is that it does not allow histopathological lesion evaluation and it uses special and expensive equipment and drugs, thereby limiting the number of capable facilities [45,[52][53][54][55][56]. For these treatments, ER is superior because it is minimally invasive and allows detailed pathological evaluation.
Regarding ER for locally recurrent ESCC after CRT, ESD was reported with an en bloc resection rate of 86-100% [15,18,19,[21][22][23], higher than that of endoscopic mucosal resection (EMR) of 46-47% [17,18,20]. The en bloc resection rate in EMR in ESCC with local recurrence after CRT is low because of the difficulty of reliable snaring due to the submucosal fibrosis [15,16]. Conversely, ESD is feasible for ESCC resection after CRT if a secure mucosal incision and submucosal dissection can be performed. However, ESD is associated with a risk of perforation in lesions with strong submucosal fibrosis. Nagami et al. reported that prior CRT was an independent predictor of lower en bloc resection rate and perforation in ESD for ESCC [57]. Additionally, it has been reported that submucosal fibrosis increases with time after CRT [58]. However, this study found no significant difference in time course after CRT; moreover, no significant correlation was found using the Pearson product-moment correlation coefficient (correlation coefficient: 0.011, p-value: 0.97). Since only a small number of cases were considered in our study, further research is required to accumulate the association between submucosal fibrosis and duration after CRT.
In this study, all lesions were associated with submucosal fibrosis in ESCC with local recurrence in the irradiated area after CRT, but en bloc resection by ESD was possible in all cases. However, the procedure was difficult to perform in lesions with strong fibrosis, and the procedure speed was reduced, suggesting the need for careful treatment techniques. Furthermore, predicting submucosal fibrosis before ESD was difficult; therefore, ESD was performed.
Several studies on the prognosis of patients who underwent ER after CRT for esophageal cancer showed a 61.9-84.2% disease-specific survival range at 18-54 months of followup after ER, and a 0-46.7% all-cause death rate [16,17,19,[21][22][23]. The 3-year overall survival (OS) was 56.1-75% and 41.6-49.1% for the 5-year OS [16,17,19,[21][22][23] (Table 3). The present study revealed no ESCC-related deaths at a median follow-up of 42 months (range: 14-117 months) after ESD, although four all-cause deaths were observed among 11 patients. The 3-year and 5-year OS rates were 90% and 72%, respectively. Therefore, our results suggest that ESD can improve the prognosis if ESCC is detected in early-stage cancer after CRT although it may be related to the fact that all patients who underwent ESD in this study had metachronous lesions, rather than residual or local recurrence.
This study has several limitations. First, it was a single-center, retrospective study with small sample size. Second, the evaluated submucosal fibrosis from endoscopic images was evaluated from stored images and written records. Pathologic fibrosis may also be underestimated because it is limited to the presence of fibrosis in the resected specimen. Third, ESD was performed by more than one endoscopist.

Conclusions
All ESCCs with metachronous recurrence after CRT were associated with moderate or high fibrosis endoscopically or histopathologically. Furthermore, ESCC after CRT reliably and safely resected en bloc by ESD but was more difficult in lesions with strong submucosal fibrosis. Thus, further prospective multicenter studies are needed to establish new evidence.  Informed Consent Statement: Informed consent was obtained for all patients in person and in writing. Consent for the study was then provided through an opt-out option on the university's website.
Data Availability Statement: Data available on request due to restrictions, e.g., privacy or ethical.