Sequence of Contact X-ray Brachytherapy (CXB) and External Beam Radiation (EBRT) in organ-preserving treatment for small rectal cancer

Original Article Sequence of Contact X-ray Brachytherapy (CXB) and External Beam Radiation (EBRT) in organ-preserving treatment for small rectal cancer Ngu Wah Than a , b , D. Mark Pritchard a , David M. Hughes c , Carrie A. Duckworth a , Helen Wong b , Muneeb Ul Haq a , b , Rajaram Sripadam b , Arthur Sun Myint a , b , * a Department of Molecular and Clinical Cancer Medicine, Institute of Systems, Molecular and Integrative Biology, The University of Liverpool, L69 3GE, United Kingdom b The Clatterbridge Cancer Centre NHS Foundation Trust, 65 Pembroke Place, Liverpool L7 8YA, United Kingdom c Department of Health Data Science, Institute of Population Health, The University of Liverpool, L7 3EA, United Kingdom ARTICLE INFO Keywords: Rectal cancer Contact X-ray brachytherapy Papillon Radiotherapy Organ preservation Watch and wait Propensity score matching Inverse probability treatment weighting ABSTRACT Background and purpose: External Beam Radiotherapy (EBRT) followed by Contact X-ray Brachytherapy (CXB) and vice versa are viable alternatives to surgery for selected rectal cancer patients who have small tumours (≤3 cm). However, the optimal sequence of treatment needs to be established. We compared two approaches using Propensity Score (PS) matching and inverse probability treatment weighting (IPTW) analyses to investigate whether the sequence of treatment affected patient outcomes. Materials and methods: This retrospective analysis (2008–2019) included patients with rectal adenocarcinoma (cT1-3,N0-1,M0, grade 1–2, size ≤3 cm) who received both EBRT and CXB, irrespective of treatment sequence. PS matching and IPTW were conducted to balance covariate standardised mean differences between groups. Oncological outcomes and rate of post-treatment rectal bleeding were assessed. Results: Following PS matching and IPTW analyses from 251 eligible patients; 103 starting with EBRT (median follow-up: 37 [IQR:18–56] months) and 148 with CXB (median follow-up: 32 [IQR:16–54] months, a significant improvement in 3-year overall survival (77% vs 85%, p =0.02, [HR:0.58 (95% CI:0.37–0.91)]) and a higher risk of post-treatment rectal bleeding (grade 1 (26%) and grade 2 (6%)) were found in patients who started with CXB (p =0.08). No significant differences were observed in local regrowth (18% vs 12%, p =0.47), distant relapse (10% vs 6%, p =0.53), 3-year organ preservation rates (70% vs 75%, p =0.20, [HR:0.66 (95% CI: 0.35–1.26)]), or disease-free survival (78% vs 82%, p =0.17, [HR: 0.47 (95% CI: 0.16–1.38)]) Conclusion: In patients with rectal cancer (≤3 cm), commencing with CXB rather than EBRT, was associated with improved overall survival, but had a higher risk of G1/2 rectal bleeding. No statistically significant differences were observed in other oncological outcomes. Introduction Colorectal cancer is the third most common cancer and the second most common cause of cancer deaths worldwide [1]. Organ-preserving treatment options for rectal cancer have been used as viable alternatives to surgery for many years, in order to avoid undesirable surgical risks and their impacts on patients’quality of life. These treatments are used particularly in patients who are frail or who have significant comorbidities and in those who are stoma averse [2–12]. The combination of Contact X-ray Brachytherapy (CXB) and External Beam Radiotherapy (EBRT) as an organ preservation approach has been shown to have proven benefits in sustaining local tumour control and maintaining long-term favourable bowel function [6–8,13–15]. Published analyses by Papillon et al., Schild et al., and Aumock et al. of patient outcomes after starting with EBRT followed by CXB have revealed 5-year local tumour control rates ranging from 71%-92%, while the disease-free survival was 60–76% for T2-3, small (≤3cm) rectal cancer cases [16–18]. However, few studies have investigated whether starting with CXB instead of EBRT results in any improvement in local tumour control or * Corresponding author at: The Clatterbridge Cancer Centre NHS Foundation Trust, 65 Pembroke Place, Liverpool L7 8YA, United Kingdom. E-mail addresses: [email protected] (N.W. Than), [email protected] (D.M. Pritchard), [email protected] (D.M. Hughes), C.A. [email protected] (C.A. Duckworth), [email protected] (H. Wong), [email protected] (M.U. Haq), [email protected] (R. Sripadam), [email protected] (A.S. Myint). Contents lists available at ScienceDirect Radiotherapy and Oncology journal homepage: www.thegreenjournal.com https://doi.org/10.1016/j.radonc.2024.110465 Received 15 February 2024; Received in revised form 18 July 2024; Accepted 25 July 2024 Radiotherapy and Oncology 199 (2024) 110465 Available online 27 July 2024 0167-8140/© 2024 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ ). reduced local recurrence, particularly for small rectal cancers (≤3 cm). The study by Dhadda et al., in which patients received CXB prior to EBRT, in the setting of post-operative treatment after local tumour excision, showed a lower risk of local tumour recurrence (4–5% vs 15%) compared to patients who received EBRT first [19]. Benezery et al. similarly found that a treatment approach involving upfront CXB for tumours ≤3 cm resulted in a more favourable initial clinical complete response (88% vs 32%) compared to EBRT first [20]. The OPERA randomised phase 3 controlled clinical trial also demonstrated an excellent 3-year organ preservation rate in the CXB first arm (97%) compared with the EBRT first arm (68%) [15]. However, in this trial, CXB was only used first for smaller tumours (≤3 cm), whereas EBRT was used initially for larger tumours (>3 cm). Hence, the results of OPERA are not fully able to inform whether one treatment sequence was preferable to the other in patients who had small tumours. Gaps in the evidence therefore remain in addressing the optimal sequence of treatment, whether to start with CXB first or EBRT, particularly for small rectal cancers (≤3 cm). Recently, a new guideline from the French National Society of Gastroenterology (SNFGE) has recommended CXB as the initial treatment for cT2N0M0 rectal cancer with a tumour diameter of ≤3 cm [21]. However, various sequences of this organ-preserving treatment are still being used internationally. We therefore aimed to compare the oncological outcomes between these two treatment approaches. We used propensity score matching and inverse probability treatment weighting (IPTW) analyses to optimise the effectiveness of the gathered observational data and ensure a balanced distribution of baseline characteristics between the two treatment groups. Materials and methods Patient selection Institutional audit approval was obtained for a retrospective review of all sequential patients who received CXB treatment at Clatterbridge Cancer Centre (CCC), Merseyside, UK from 2008 to 2019. Data from patients who had undergone both EBRT and CXB with curative intent, regardless of the treatment sequence, were collected from the database system including demographics, tumour size and stage, treatment details and outcome data. Only patients who had well to moderately differentiated localised rectal adenocarcinomas (cT1-3, cN0-1, cM0), small tumour diameters (≤3 cm), and no features of extramural venous invasion were included in the current analysis. The tumour diameter was determined from pre-treatment MRI scan reports, digital rectal examinations (DRE) performed by the oncologist or colorectal surgeon, and the CXB applicator size that was used during the first treatment session. Patient management Patients were referred to CCC by local Multidisciplinary Teams (MDTs) at colorectal cancer centres throughout the UK. In some cases, patients had already received EBRT locally and were referred for consideration of CXB because they had not had a complete response to this treatment. Other patients were referred de novo prior to receiving any therapy and in some of these, it was decided to administer CXB first followed by EBRT. The decision to begin treatment with either (chemo) radiation or refer patients for consideration of starting with CXB was based on the preference of the local MDT, as there are only four CXB treatment centres in the UK. EBRT was administered either as chemoradiation (45–50.4 Gy/25 #/35 days), long-course (40–45 Gy/20#/28 days), or short-course (25 Gy/5#/5 days) regimens. The decision to choose a particular regimen was based on the patient’s performance status and comorbidities. Conformal 3-dimentional radiation was the primary technique before 2013 with a shift to Intensity Modulated Radiation Therapy (IMRT) thereafter. Radiation volume was typically limited to treat the mesorectum and generally did not go above S2/3. No adjuvant chemotherapy was routinely administered after radiotherapy treatment. CXB was performed on an outpatient basis using the Papillon-50 machine (50kVp X-rays (HVL 0.64 Al, 2.7 mA), Ariane, Alfreton, UK), 20-30 Gy per fraction delivered 2 weeks apart, through a rectal treatment applicator (size 30, 25, or 22 mm) at a focal source surface distance of 29, 32, or 38 mm respectively. A total dose of 90–110 Gy (rectal mucosal surface dose) was delivered in 3–4 fractions over 4–6 weeks. Patients who achieved a clinical/near complete response 8–12 weeks after completion of all treatments were assessed regularly with a DRE, flexible sigmoidoscopy/rigid rectoscopy alternately at their local centre by the colorectal surgeon and by the radiation oncologist at CCC except for patients who were unable to attend regular follow-ups at the CCC, and MRI scans at 12-week intervals at their local centre in the first 2 years and every 6 months in the 3rd year. Usually, only endoscopic examination was done in the 4th and 5th years. Patients who did not achieve a complete response were referred for salvage surgery if they were fit enough for this. Outcome measures The primary outcome measures were overall survival (OS) and disease-free survival (DFS). Disease-free survival was calculated from the date of the last radiotherapy treatment to the date of locoregional recurrence after R0/1 resection of the primary tumour, non-salvageable local regrowth/R2 salvage resection, the occurrence of second primary, distant metastasis, or last follow-up. Overall survival was defined as the period from the last radiotherapy treatment to the date of the last data review or death from any cause. Secondary outcome measures included the rate of initial clinical Complete Response (cCR), local regrowth, nodal/regional relapse, distant relapse, organ preservation rate, and radiation toxicities, specifically post-treatment rectal bleeding. Statistical analysis We utilised two propensity score methods to minimise the standardized mean differences in covariates, thereby achieving betterbalanced cohorts in this retrospective single-centre analysis. Typically, the propensity score matching is calculated through multivariate logistic regression forming a matched cohort. However, matching often results in a reduced sample size, leading to biased estimation in survival analysis in small samples with low event rates. To overcome this, we also employed IPTW analysis which maintains information from most patients by creating a synthetic sample, in which the distribution of measured baseline covariates is independent of treatment assignment [22,23]. Firstly, the association between radiotherapy regimen and individual time-to-event outcome (OS and DFS) was examined using survival model in an unadjusted analysis. Secondly, considering the potential influence of various baseline covariates, the second set of survival models adjusted for these covariates for each outcome. The variables of age, gender, fitness for surgery, performance status, tumour (T) stage, nodal (N) stage, EBRT regimen, CXB total dose, and overall treatment time were considered possible confounders. The overall treatment time was calculated from the start date to the end date of all radiotherapy treatments, including any gaps between therapies. This duration was then divided into two categories. Standard treatment time was defined as ≤12 weeks (EBRT =1week, maximum gap time =4 weeks, total CXB treatment time =4–6 weeks) for patients who received short-course external beam radiation or ≤25 weeks (EBRT =5 weeks, maximum gap time =14 weeks, total CXB treatment time =4–6 weeks) for those who received long-course (chemo)radiation. Unconventional treatment time was designated as ≥12 weeks for patients who received shortcourse external beam radiation or ≥25 weeks for those who received long-course (chemo)radiation. N.W. Than et al. Radiotherapy and Oncology 199 (2024) 110465 2 After that, we adopted two propensity score weighting methods to minimise the influence of confounding variables in our retrospective, observational study. The propensity score indicates the probability of an individual receiving a specific treatment, as determined by predefined covariates. This probability is calculated using a logistic regression model in which treatment is treated as a binary dependent variable, and a designated set of covariates serves as independent variables to ensure balance. The propensity scores were used in two ways. First, we performed propensity score matching using the nearest-neighbour method with a calliper of 0.25. Then, IPTW was calculated based on the propensity scores to weight the full cohort. The balance of covariates before and after propensity matching and IPTW was assessed by examining the standardised mean difference (SMD) between the groups. An SMD of greater than 0.1 is considered a significant imbalance. Survival analysis methods were used to evaluate the two time-to-event outcomes. Visual inspection of survival differences using Kaplan-Meier curves and statistical assessments using Log-rank tests were performed. The association between treatment type and survival risks was analysed using Cox proportional hazards models, for each time-to-event outcome, the proportional hazards assumption required by the Cox model was assessed by Schoenfeld residuals. P-value <0.05 was considered statistically significant. Statistical analyses were performed in R 4.3.0. Results A total of 251 eligible patients, who received both EBRT and CXB, regardless of the sequence, with curative intent were included in the study. One hundred and three patients received EBRT and 148 had CXB as their initial treatment. The median follow-up was 37 [IQR:18–56] months for the EBRT-first group and 32 [IQR:16–54] months for the CXB-first group. The two groups were reasonably matched in terms of age and gender profiles, performance status and fitness for surgery. However, significant covariate imbalances in terms of patients’tumour (T) stage, nodal (N) stage, EBRT regimens, and prolonged overall treatment time were evident between the two groups (Table 1). Patients in the EBRT first group had significantly more advanced tumour stages and were also more likely to have received concurrent chemotherapy as part of their EBRT regimen. Propensity score matching and IPTW were therefore performed to consider these differences and to allow a more appropriate statistical analysis. Following PS matching and IPTW, there was a good balance across all covariates, with no statistically significant differences remaining apart from slight deviations of SMD observed for the distance from the anal verge in the propensity-matched model, and for fitness for surgery and EBRT regimens in the IPTW model (Supplementary Fig. 1A and 1B). The demographic data for all cohorts are summarised in Table 1. In the unadjusted analysis, the rates of initial cCR 79% vs 88%, (p = 0.07), local regrowth 19% vs 15%, (p =0.81), nodal/regional relapse 4% vs 2.7%, (p =0.60), and distant relapse 13% vs 8%, (p =0.24) were observed in the EBRT-first and CXB-first groups, respectively (Table 2). The 3-year organ preservation rates were 69% (95%CI: 55–78) and 73% (95%CI: 63–79) (p =0.42, HR: 0.82 [95%CI: 0.51–1.32]) (Fig. 2A) with 3-year disease-free survival rates of 78% (95%CI: 72–82) and 80% (95% Table 1 Baseline characteristics of unadjusted, propensity-score matched, and IPTW cohorts. Characteristics Unadjusted Propensity score-matched IPTW EBRT first n=103(%) CXB first n=148(%) P value EBRT first n=72(%) CXB first n=72(%) P value EBRT first n=290(%) CXB first n=243(%) P value Mean age 71.9 ±10.8 73.7 ±11.2 0.22 71.6 ±11.0 69.3 ±11.8 0.24 72.1 ±9.3 72.9 ±11.0 0.52 Gender Male Female 75(73) 28(27) 101(68) 47(32) 0.44 54(75) 18(25) 50(69) 22(31) 0.58 203(71) 87(29) 165(68) 78(32) 0.81 Fitness for surgery Fit Unfit 75(73) 28(27) 101(68) 47(32) 0.44 32(44) 40(56) 37(51) 35(49) 0.51 154(53) 136(47) 121.5(50) 121.5(50) 0.39 WHO performance status 1–2 2–3 63(61) 40(39) 78(53) 70(47) 0.18 46(64) 26(36) 47(65) 25(35) 1.00 180(62) 110(38) 124(51) 119(49) 0.82 T stage T1 T2 T3 3(3) 51(49) 49(48) 15(10) 104(70) 29(20) <0.001 3(4) 44(61) 25(35) 4(6) 46(64) 22(31) 0.83 14(5) 189(65) 87(30) 17(7) 158(65) 68(28) 0.82 N stage N0 N1 67(65) 36(35) 128(87) 20(13) <0.001 56(78) 16(22) 56(78) 16(22) 1.0 235(81) 55(19) 190(78) 53(22) 0.70 Distance from anal verge <6cm ≥6–12 cm 65(63) 38(37) 105(71) 43(29) 0.19 45(63) 27(37) 56(78) 16(22) 0.69 220(76) 70(24) 175(72) 68(28) 0.53 EBRT regimen Short course Long course Chemoradiation 21(20) 21(20) 61(60) 82(55) 13(9) 53(36) <0.001 20(28) 15(21) 37(51) 22(31) 9(12) 41(57) 0.41 114(45) 33(13) 106(42) 111(45) 32(13) 104(42) 0.96 CXB dose 80-90gy 110gy 78 (76) 25(24) 111 (75) 37(25) 1.0 57(79) 15(21) 52(72) 20(28) 0.44 218(75) 72(25) 180(74) 63(26) 0.77 Overall treatment time Standard Unconventional 67(65) 36(35) 123(83) 25(17) 0.002 56(78) 16(22) 56(78) 16(22) 1.00 232(80) 58(20) 190(78) 53(22) 0.71 Table 2 Oncological outcomes of different analyses. Outcome Unadjusted Propensity score-matched IPTW EBRT first n=103(%) CXB first n=148(%) P value EBRT first n=72(%) CXB first n=72(%) P value EBRT first n=290(%) CXB first n=243(%) P value Initial response 81(79) 130(88) 0.07 46 (78) 54(85) 0.39 249(86) 214(88) 0.57 Local regrowth 15(19) 20(15) 0.81 10 (18) 7 (12) 0.47 55(19) 24(10) 0.20 Nodal/regional relapse 4(4) 4(2.7) 0.60 2(3) 1(1.4) 1.00 9(3) 8(3) 1.00 Distant relapse 13(13) 12(8) 0.24 7(10) 4(6) 0.53 23(8) 17(7) 0.75 Absolute organ preservation rate 71(69) 111(75) 0.29 49 (68) 56 (78) 0.26 212 (73) 197(81) 0.26 Rectal bleeding G1 G2 18(18) 15(15) 3(3) 48(32) 38 (26) 10(6) 0.01 13 (18) 23 (32) 0.08 38(13) 90(37) 0.004 N.W. Than et al. Radiotherapy and Oncology 199 (2024) 110465 3 CI: 75–85). Cox proportional hazard model of disease-free survival showed no significant difference between the two regimens in both univariable (p =0.68, HR: 0.87 [95%CI: 0.45–1.67]) and multivariable analyses (p =0.66, HR: 0.84 [95%CI: 0.39–1.82]) (Fig. 3A). The 3-year overall survival rates were 78% (95%CI: 73–84) and 79% (95% CI: 74–84) and no statistical difference in hazard ratio was observed between the two groups in univariable analysis (p =0.44, HR: 0.88 [95% CI: 0.63–1.22]). However, the multivariable analysis indicated a Fig. 1. Flow chart illustrating overall clinical outcomes for the entire study group. N.W. Than et al. Radiotherapy and Oncology 199 (2024) 110465 4 significant difference, suggesting that patients receiving CXB first were more likely to survive (p =0.03, HR: 0.65 [95%CI: 0.43–0.97]) (Fig. 4A). Post-treatment G1-2 self-limiting rectal bleeding occurred in 18% of EBRT-first patients and 32% of CXB-first patients (p =0.01). The overall clinical outcomes for the entire study group are illustrated in Fig. 1. In the propensity-matched cohort, the initial cCR rate was 78% vs 85%, (p =0.39) with the local regrowth rate of 18% vs 12%, (p =0.47), nodal/regional relapse rate of 3% vs 1.4%, (p =1.00), and distant relapse rate of 10% vs 6%, (p =0.53) for the EBRT-first and CXB-first regimens respectively (Table 2). The 3-year organ preservation rates were 70% (95%CI: 55–80) and 75% (95%CI: 72–85), (p =0.20, HR: 0.66 [95%CI: 0.35–1.26]) (Fig. 2B) with 3-year disease-free survival rates of 78% (95%CI: 72–90) and 82% (95%CI: 78–97). The Cox model of disease-free survival indicated no significant difference between groups in both univariable (p =0.17, HR: 0.47 [95%CI: 0.16–1.38]) and multivariable analyses (p =0.23, HR: 0.50 [95%CI: 0.16–1.55]) (Fig. 3B). The 3-year overall survival rates were 77% (95%CI: 73–83) and 85% (95%CI: 80–95). Significant differences in hazard ratios between the two groups were observed in both univariable (p =0.02, HR: 0.58 [95%CI: 0.37–0.91]) and multivariable (p =0.001, HR: 0.44 [95% CI: 0.28–0.73]) analyses, indicating that patients receiving CXB first were more likely to survive (Fig. 4B). Post-treatment rectal bleeding occurred in 18 % of EBRT-first patients and 32 % of CXB-first patients (p =0.08). The IPTW analysis revealed the following outcomes for EBRT-first and CXB-first treatments, respectively: initial cCR rates (86% vs. 88%, p=0.57), local regrowth rates (19% vs. 10%, p =0.20), similar nodal/ regional rates (both 3%, p =1.0), and distant relapse rates (8% vs. 7%, p =0.75) as shown in Table 2. The 3-year organ preservation rates were 73% (95%CI: 55–80) and 80% (95%CI: 72–90), (p =0.27, HR: 0.47 [95%CI: 0.35–1.10]), with 3-year disease-free survival of 87% (95 %CI: 76–92) and 88% (95%CI: 78–95), and overall survival of 78% (95%CI: 65–85) and 78% (95%CI: 70–82). Cox proportional hazard ratios of Fig. 2A. Kaplan-Meier curve comparing organ-preservation rates between two sequences of treatment in unadjusted model. Fig. 2B. Kaplan-Meier curve comparing organ-preservation rates between two sequences of treatment in propensity-matched model. N.W. Than et al. Radiotherapy and Oncology 199 (2024) 110465 5 disease-free survival indicated no significant difference between the two regimens in both univariable (p =0.20, HR: 0.46 [95%CI: 0.16–1.51]) and multivariable analyses (p =0.95, HR: 1.03 [95%CI: 0.46–2.29]). Despite no statistical significance (p =0.06, HR: 0.70 [95%CI: 0.48–1.03]) being observed in univariable analysis, the multivariable analysis demonstrated a significant improvement in overall survival with CXB-first treatment (p =0.005, HR: 0.58 [95%CI: 0.40–0.85]). A higher risk of post-treatment rectal bleeding was also seen in the CXBfirst arm (p =0.004) (Supplementary Table 1). Discussion Although commencing organ-preserving treatment with CXB followed by EBRT for small rectal tumours (≤3 cm) has previously suggested that this therapy sequence results in improved local tumour control rates [15,20], the optimal treatment selection for small rectal cancers has not yet been established. It is therefore imperative to investigate whether delivering CXB first has any benefits over the administration of EBRT as the initial treatment, particularly in terms of long-term clinical outcomes. The patient and tumour characteristics presented in this study closely resembled those outlined in the OPERA trial and Benezery et al. studies for the CXB first group, except for a somewhat lower performance status (WHO 2–3) (44% vs. 9–10%) and a higher risk for surgery (57%). In addition, nearly half of the patients (41%) in our study underwent a short course of EBRT, whereas almost all patients in the other studies received a long course of chemoradiation [15,20]. In the current study, patients who underwent CXB as their initial treatment exhibited an initial response rate of 85–88% which is very similar to the findings presented in Benezery et al. (88%). The rates of local regrowth (10–15%) and distant relapse (6–8%) were also comparable to those studies. The 3-year organ preservation rate was 73–80%, whereas the OPERA trial reported 97% organ preservation at 3 years [15]. A higher risk of post-treatment occasional self-limiting grade 1–2 Fig. 3A. Kaplan-Meier curve comparing disease-free survival between two sequences of treatment in unadjusted model. Fig. 3B. Kaplan-Meier curve comparing disease-free survival between two sequences of treatment in propensity-matched model. N.W. Than et al. Radiotherapy and Oncology 199 (2024) 110465 6 rectal bleeding was also seen in the CXB first group, consistent with the findings of the OPERA study. This might be related to a higher incidence of patients who were on anticoagulant therapy in the CXB-first group (23% vs 14%). For the patients in this cohort who started with EBRT, the initial cCR rate was 78–86%, the local regrowth rate was 18–19%, and the distant relapse rate was 8–13%. The 3-year organ preservation rate was 69–73% with a 3-year disease-free survival rate of 78–87%. These findings are comparable to the outcomes in the published studies by Papillon et al., Schild et al., and Aumock et al. [16–18]. We found a significant improvement in overall survival among patients who underwent CXB as their initial treatment (79–85% at 3 years), compared to patients who received EBRT first (77–78% at 3 years), despite no significant difference being observed in disease-free survival (80–88% vs 78–87% at 3 years). One possible explanation could be that, despite not reaching statistical significance, the CXB-first arm exhibited lower rates of local regrowth and distant relapse, along with a higher rate of organ preservation, in both unadjusted and propensity-matched cohorts. Additionally, some patients who underwent salvage surgery after failing organ-preserving treatment were only followed up at their local centres and no longer had appointments at CCC. Consequently, the disease-free survival endpoint, marked by the last follow-up date recorded in the CCC database, appeared to be shorter. However, the status of patients as being alive/deceased was continuously updated in the database system and is therefore more accurate. These factors could have contributed to the observed shorter disease-free survival period while potentially extending the overall survival. Despite conducting propensity-matching analyses, our study still retains several potential confounders which may have potentially affected the results. The major limitation is the inherent nature of a nonrandomized retrospective observational study, which cannot completely address unobserved differences between the groups. The organ preservation rate that we found may have been influenced by variations in salvage surgery for residual/local regrowth between the Fig. 4A. Kaplan-Meier curve comparing overall survival between two sequences of treatment in unadjusted model. Fig. 4B. Kaplan-Meier curve comparing overall survival between two sequences of treatment in propensity-matched model. N.W. Than et al. Radiotherapy and Oncology 199 (2024) 110465 7 two groups. Salvage local excision (TEMS or TAMIS) was performed in 5 out of 37 patients in the EBRT-first group, whereas only 3 out of 38 patients in the CXB-first group underwent local excision (Fig. 1). Additionally, more than half of the patients from both cohorts were frail and were at high risk for surgery, so unfortunately, they were not fit enough to undergo salvage surgery for their residual/ recurrent diseases. This may potentially have contributed to the observed reduction in diseasefree survival despite the more prolonged overall survival. It is also important to note there was heterogeneity in the treatment regimens that were used within the CXB-first group: 55 patients received a complete course of CXB followed by EBRT whereas 93 patients underwent a split course of CXB and EBRT (1–3 fractions of CXB followed by EBRT and then another 1–2 fractions of CXB for any residual disease). Patients who received a split course of treatment tended to exhibit a more limited initial response to CXB treatment, leading to the administration of this specific regimen. This variability may have introduced complexities in assessing the true outcomes of this treatment approach. Another important caveat of our study relates to the initial method of patient referral. It is known that approximately 30% of rectal cancer patients achieve a clinical complete response following initial neoadjuvant (chemo)radiation [24]. Typically, these patients are monitored through active surveillance and are not referred for additional treatment unless they experience local/distant disease relapse. It is therefore probable that some patients who would otherwise have been eligible for this study, experienced a sustained complete clinical response following EBRT at their local treatment centre and were therefore not referred to CCC for further evaluation. Conclusions Despite these limitations, our study suggests that commencing treatment for small rectal cancer (≤3 cm) with CXB, as opposed to EBRT, appears to be associated with increased overall survival, despite a higher incidence of occasional self-limiting grade 1–2 rectal bleeding. These findings align, in part, with other emerging evidence from recent studies. We therefore recommend that adequately powered prospective clinical trials are performed in order to definitively establish whether CXB-first provides benefits over EBRT-first in small rectal cancers. CRediT authorship contribution statement Ngu Wah Than: Writing –original draft, Visualization, Software, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. D. Mark Pritchard: Writing –review &editing, Validation, Supervision, Project administration, Methodology, Funding acquisition, Data curation, Conceptualization. David M. Hughes: Validation, Supervision, Software, Methodology, Data curation, Conceptualization. Carrie A. Duckworth: Supervision, Methodology, Conceptualization. Helen Wong: Resources, Data curation. Muneeb Ul Haq: Data curation. Rajaram Sripadam: Resources, Data curation. Arthur Sun Myint: Writing –review &editing, Supervision, Resources, Project administration, Methodology, Funding acquisition, Data curation, Conceptualization. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgment This project was funded by the European Union’s Horizon 2020 research and innovation programme, Marie Skłodowska-Curie grant agreement number 857894—CAST. Appendix A. Supplementary material Supplementary data to this article can be found online at https://doi. org/10.1016/j.radonc.2024.110465. References [1] Cancer I.A.f.R.o. 10_8_9-Colorectum-fact-sheet; 2020. <https://gco.iarc.fr/today/ data/factsheets/cancers/10_8_9-Colorectum-fact-sheet.pdf/>[accessed 12 November 2023]. [2] Habr-Gama A, S˜ ao Juli˜ ao GP, Vailati BB, et al. Organ preservation in cT2N0 rectal cancer after neoadjuvant chemoradiation therapy: the impact of radiation therapy dose-escalation and consolidation chemotherapy. Ann Surg 2019;269:102–7. https://doi.org/10.1097/SLA.0000000000002447. [3] Renehan AG, Malcomson L, Emsley R, et al. Watch-and-wait approach versus surgical resection after chemoradiotherapy for patients with rectal cancer (the OnCoRe project): a propensity-score matched cohort analysis. Lancet Oncol 2016; 17:174–83. https://doi.org/10.1016/S1470-2045(15)00467-2. [4] van der Valk MJM, Hilling DE, Bastiaannet E, et al. Long-term outcomes of clinical complete responders after neoadjuvant treatment for rectal cancer in the International Watch &Wait Database (IWWD): an international multicentre registry study. Lancet 2018;391:2537–45. https://doi.org/10.1016/S0140-6736 (18)31078-X. [5] Papillon J. Present status of radiation therapy in the conservative management of rectal cancer. Radiother Oncol 1990;17:275–83. https://doi.org/10.1016/01678140(90)90001-d. [6] Gerard JP, Chapet O, Nemoz C, et al. Improved sphincter preservation in low rectal cancer with high-dose preoperative radiotherapy: the Lyon R96–02 randomized trial. J Clin Oncol 2004;22:2404–9. https://doi.org/10.1200/JCO.2004.08.170. [7] Appelt AL, Pløen J, Harling H, et al. High-dose chemoradiotherapy and watchful waiting for distal rectal cancer: a prospective observational study. Lancet Oncol 2015;16:919–27. https://doi.org/10.1016/S1470-2045(15)00120-5. [8] Sun Myint A, Smith FM, Gollins S, et al. Dose escalation using contact X-ray brachytherapy after external beam radiotherapy as nonsurgical treatment option for rectal cancer: outcomes from a single-center experience. Int J Radiat Oncol Biol Phys 2018;100:565–73. https://doi.org/10.1016/j.ijrobp.2017.10.022. [9] Couwenberg AM, Burbach JPM, van Grevenstein WMU, et al. Effect of neoadjuvant therapy and rectal surgery on health-related quality of life in patients with rectal cancer during the first 2 years after diagnosis. Clin Colorectal Cancer 2018;17: e499–512. https://doi.org/10.1016/j.clcc.2018.03.009. [10] Juul T, Ahlberg M, Biondo S, et al. Low anterior resection syndrome and quality of life: an international multicenter study. Dis Colon Rectum 2014;57:585–91. https://doi.org/10.1097/DCR.0000000000000116. [11] Morris EJ, Taylor EF, Thomas JD, et al. Thirty-day postoperative mortality after colorectal cancer surgery in England. Gut 2011;60:806–13. https://doi.org/ 10.1136/gut.2010.232181. [12] Manceau G, Karoui M, Werner A, Mortensen NJ, Hannoun L. Comparative outcomes of rectal cancer surgery between elderly and non-elderly patients: a systematic review. Lancet Oncol 2012;13:e525–36. https://doi.org/10.1016/ S1470-2045(12)70378-9. [13] Papillon J. Intracavitary irradiation of early rectal cancer for cure A series of 186 cases. Cancer 1975;36:696–701. https://doi.org/10.1002/1097-0142(197508)36: 2+<696::aid-cncr2820360813>3.0.co;2-x. [14] G´ erard JP, Barbet N, Gal J, et al. Planned organ preservation for early T2–3 rectal adenocarcinoma: a French, multicentre study. Eur J Cancer 2019;108:1–16. https://doi.org/10.1016/j.ejca.2018.11.022. [15] Gerard JP, Barbet N, Schiappa R, et al. Neoadjuvant chemoradiotherapy with radiation dose escalation with contact x-ray brachytherapy boost or external beam radiotherapy boost for organ preservation in early cT2–cT3 rectal adenocarcinoma (OPERA): a phase 3, randomised controlled trial. Lancet Gastroenterol Hepatol 2023;8:356–67. https://doi.org/10.1016/S2468-1253(22)00392-2. [16] Papillon J, Berard P. Endocavitary irradiation in the conservative treatment of adenocarcinoma of the low rectum. World J Surg 1992;16:451–7. https://doi.org/ 10.1007/BF02104446. [17] Schild SE, Martenson JA, Gunderson LL. Endocavitaryradiotherapy of rectal cancer. Int J Radiat Oncol Biol Phys 1996;34:677–82. https://doi.org/10.1016/ 0360-3016(95)02098-5. [18] Aumock A, Birnbaum EH, Fleshman JW, et al. Treatment of rectal adenocarcinoma with endocavitary and external beam radiotherapy: Results for 199 patients with localized tumors. Int J Radiat Oncol Biol Phys 2001;51:363–70. https://doi.org/ 10.1016/s0360-3016(01)01677-7. [19] Dhadda A, Sun Myint A, Thamphya B, et al. A multi-centre analysis of adjuvant contact X-ray brachytherapy (CXB) in rectal cancer patients treated with local excision–Preliminary results of the CONTEM1 study. Radiother Oncol 2021;162: 195–201. https://doi.org/10.1016/j.radonc.2021.07.021. [20] Benezery K, Montagne L, Evesque L, et al. Clinical response assessment after contact X-Ray brachytherapy and chemoradiotherapy for organ preservation in rectal cancer T2–T3 M0: The time/dose factor influence. Clin Transl Radiat Oncol 2020;24:92–8. https://doi.org/10.1016/j.ctro.2020.07.001. [21] Gastroenterology, F.N.S.o. Diseases and cancers of the digestive system. Rectal Cancer;2023. <https://www.snfge.org/tncd/cancer-du-rectum/>[accessed 5 September 2023]. N.W. Than et al. Radiotherapy and Oncology 199 (2024) 110465 8 [22] Austin PC, Fine JP. The use of propensity score methods with survival or time-toevent outcomes: reporting measures of effect similar to those used in randomized experiments. Stat Med 2014;33:1242–58. https://doi.org/10.1002/sim.8008. [23] Chesnaye NC, Stel VS, Tripepi G, et al. An introduction to inverse probability of treatment weighting in observational research. Clin Kidney J 2022;15:14–20. https://doi.org/10.1093/ckj/sfab158. [24] Glynne-Jones R, Hughes R. Complete response after chemoradiotherapy in rectal cancer (watch-and-wait): have we cracked the code? Clin Oncol (R Coll Radiol) 2016;28:152–60. https://doi.org/10.1016/j.clon.2015.10.011. N.W. Than et al. Radiotherapy and Oncology 199 (2024) 110465 9