Assessment of circulating tumor DNA in patients with locally advanced rectal cancer treated with neoadjuvant therapy Authors: Chiara Molinari, Giorgia Marisi, George Laliotis, Erik Spickard, Ilario Giovanni Rapposelli, Elisabetta Petracci, Giby V George, Punashi Dutta, Shruti Sharma, Meenakshi Malhotra, Andrea Prochowski Iamurri, Giacomo Feliciani, Minetta C Liu, Paola Ulivi, Matteo Canale, Luca Saragoni, Graziana Gallo, Giovanni Luca Frassineti, Margherita Muratore, Antonino Romeo, Adham Jurdi, Giovanni Martinelli, Alessandro Passardi Published: 2024 ABSTRACT AbstractPost-neoadjuvant therapy (post-NAT) and post-surgical circulating tumor DNA (ctDNA) risk stratification may enhance the management of patients with locally advanced rectal cancer (LARC). In this study, we assessed the prognostic value of ctDNA-based MRD detection in LARC patients using a personalized, tumor-informed ctDNA assay. Plasma samples from LARC patients (N = 30) were analyzed retrospectively using the Signatera™ assay. The neoadjuvant rectal (NAR) score was calculated and compared to ctDNA status to predict recurrence risk and survival outcomes. ctDNA-positive patients post-NAT and post-surgery had worse Disease Free Survival (DFS) (HR: 7.82; p = 0.001, HR: 19.65; p = 0.001) when compared to ctDNA-negative patients. In the post-NAT setting, patients who responded to NAT had superior DFS compared to patients who did not clear their ctDNA or showed no radiological response (HR: 24.7, p = 0.001 and HR: 5.1, p = 0.054, respectively). When ctDNA status is used alongside the NAR score in the post-NAT setting, patients who were ctDNA-positive with an intermediate or high NAR score showed significantly worse DFS (HR: 47.5, p < 0.001) compared to ctDNA-negative patients with either a low or intermediate/high NAR score (HR: 9.8, p = 0.0301). Post-NAT ctDNA status, whether used alone or in combination with the NAR score, may predict NAT response, and improve risk stratification.Keywords: Circulating tumor DNA, Locally advanced rectal cancer, Biomarker, Neoadjuvant therapy, NAR scoreSubject terms: Cancer, Immunology, Molecular biology, Gastroenterology, Molecular medicine, Oncology Introduction Presently, colorectal cancer (CRC) is the third most common cancer worldwide and the second-leading cause of cancer-related death1. Approximately 30% of CRCs arise from the rectum with surgical resection being the mainstay of treatment2. For microsatellite stable (MSS) rectal cancers with high-risk pathological features or cases of locally advanced disease, multimodality treatment is recommended in the form of either neoadjuvant chemoradiation followed by total mesorectal excision (TME) and adjuvant chemotherapy or, more recently, preoperative total neoadjuvant therapy (TNT) followed by TME3–5. TNT improves the likelihood of achieving pathological complete response (pCR), which is expected to lead to lower rates of local recurrence, and distant metastasis, and thus improved overall survival6. Clinical complete response (cCR), evaluated post-neoadjuvant therapy (post-NAT) based on digital rectal examination (DRE), rectal magnetic resonance imaging (MRI) findings, and endoscopic evaluation is often assessed as a preemptive surrogate for pCR; as such, patients who exhibit cCR may be deemed appropriate for a non-operative management (NOM) strategy to obviate the need for surgical intervention2. However, there is often little concordance between cCR and pCR2. Moreover, although pCR remains a significant prognostic factor, it is a poor surrogate endpoint for overall survival7,8. Post-NAT risk stratification and appropriate decision-making may improve outcomes in locally advanced rectal cancer (LARC). Ideally, a substantial proportion of patients deemed to have low-risk disease can safely forgo additional treatment or be considered for therapeutic de-escalation9. Accordingly, the neoadjuvant rectal (NAR) score was first developed to serve as a surrogate endpoint in clinical trials following TNT10. It has been shown that the NAR score, calculated using the clinical tumor stage and pathological nodal and tumor stage, can reliably be used as a composite, shorter-term assessment of neoadjuvant treatment response10,11. More recently the radiological (r)NAR score obtained before surgery has been shown to be a stage-independent predictor of recurrence-free and overall survival in treated LARC patients12. Circulating tumor DNA (ctDNA) has been validated as a prognostic and predictive biomarker, which enables the detection of molecular residual disease (MRD) following surgical resection, identifying patients with a high risk of recurrence who may benefit from adjuvant therapy. Following curative intent therapy, ctDNA analysis in the surveillance setting may enable the identification of early recurrence or disease progression. ctDNA has thus emerged as a promising MRD assessment tool across various malignancies13–16. In fact, the utility of ctDNA analysis after TNT and/or surgery has been demonstrated in several studies with LARC2,17–22. Here, we evaluated the prognostic role of ctDNA-based MRD detection in patients with LARC using a personalized, tumor-informed ctDNA assay (Signatera™). We report on pre- and post-treatment ctDNA positivity rates and the association with response and disease-free survival (DFS), which could impact further treatment decisions. We propose that post-neoadjuvant ctDNA analysis either as a sole surrogate outcome measure or as an adjunct to currently used response assessment methods may improve patient outcomes in rectal cancer. Furthermore, we provide data that post-NAT ctDNA status is associated with pCR and thus, may be a more reliable indicator of NAT response than the composite and often subjective cCR assessment. Results Patient cohort Of the initial collected case series of 33 eligible patients, three cases failed Whole Exome Sequencing (WES) generation due to insufficient tumor content (N = 2) or failed tumor-germline concordance (N = 1). Therefore, 30 patients (plasma time points, n = 90) (stage II disease [N = 7], and stage III disease [N = 23]) with complete ctDNA (plasma time points, n = 90) and annotated clinical data were included (Table 1; Fig. 1). The median age of the cohort was 67 years (IQ-IIIQ: 63–76), with 70% (21/30) being males. Additional patient characteristics are listed in Table 1. Table 1Patient demographics and clinical information.CharacteristicN = 301Gender Male21 (70%) Female9 (30%)Age (years)67 (63–76)Clinical stage II7 (23%) III23 (77%)Number of neoadjuvant cycles5 (4–6)Neoadjuvant treatment regimen Fluorouracil + Oxaliplatin (FOLFOX)13 (43%) Capecitabine17 (57%)Radiological response Non-responders5 (17%) Responders25 (83%)Radiological neoadjuvant rectal score13.8 (8.43–30.07)Neoadjuvant rectal score11.7 (0–50.36)Surgery type Rectal anterior resection (RAR)26 (87%) Abdominoperineal resection (APR)4 (13%)Pathological complete response No24 (80%) Yes6 (20%)Adjuvant treatment No adjuvant8 (27%) Adjuvant22 (73%)Adjuvant treatment regimen Fluorouracil + Oxaliplatin (FOLFOX)12 (55%) Capecitabine + Oxaliplatin (CAPOX)6 (27%) Capecitabine4 (18%)Radiological recurrence No recurrence22 (73%) Recurrence8 (27%)Recurrence site Lung2 (25%) Lung, liver2 (25%) Liver1 (13%) Lung, liver, bone1 (13%) Bone, adrenal1 (13%) Pelvis1 (13%)Vital status Alive20 (67%) Deceased10 (33%)Recurrence-free survival (months)254 (62.7-NR)Total follow-up (months)260 (92.7-NR)1n (%); Median (IQ–IIIQ).Median (95% CI). Patient demographics and clinical information. 1n (%); Median (IQ–IIIQ). Median (95% CI). (A) Overview plot showing longitudinal ctDNA status, treatment regimen, and clinical outcomes for patients with stage II–III rectal cancer. (B) Heat map showing clinicopathological and ctDNA detection data for patients. ACT adjuvant chemotherapy, ctDNA circulating tumor DNA, NAR neoadjuvant rectal score, NAT neoadjuvant therapy, pCR pathologic complete response. ctDNA detection rates at pre- and post-treatment time points and association with patient outcomes At baseline (post-diagnostic biopsy and pre-NAT), ctDNA detection rate was 90% (27/30). Specifically, ctDNA positivity was seen in 100% (7/7) of stage II and 87% (20/23) of stage III patients (Fig. 1). All 30 patients within our annotated cohort received NAT. Thirteen patients (43%) were treated with sequential short course radiotherapy (25 Gy in 5 daily fractions) and chemotherapy (FOLFOX4 regimen) whereas 17 patients (57%) were treated with long course radiotherapy (45–50.4 Gy in 25–28 daily fractions) concomitant to continuous oral administration of capecitabine (at the dose of 825 mg/m2 twice daily for 7 days/week). Post-NAT and prior to surgery (week 11), ctDNA positivity was observed in 5 patients (4 of those with stage III disease) all with stage III disease and 4 of these were ctDNA-positive at baseline. All post-NAT ctDNA-positive patients (N = 5) showed disease recurrence (100%) compared to only 28% (7/25) of ctDNA-negative patients, correlating with a marked reduction in DFS (HR: 7.82, 95% CI 2.28–26.82, p < 0.001) (Fig. 2A). The median DFS for ctDNA-positive patients was 13.2 months, compared to not reached (NR) for ctDNA-negative patients. Of the 27 patients who were ctDNA-positive at baseline, 23 cleared their ctDNA post-NAT; only 8.7% (2/23) were found to be non-responders by radiological imaging whereas 50% (2/4) of persistently ctDNA-positive patients were deemed to be radiological non-responders (Fig. 2B). On evaluating both radiological response as well as ctDNA status post-NAT, patients who responded to NAT (evidenced by negative ctDNA and radiological status) were found to have superior DFS compared to patients who were either unable to clear their ctDNA or did not show a radiological response (HR: 24.7, 95% CI 3.58–171; p = 0.001 and HR: 5.1, 95% CI 0.97–26; p = 0.054) (Fig. 2B). Additionally, pts who were ctDNA-positive with an intermediate/high rNAR score had significantly worsened DFS (HR: 14.12, 95% Cl 1.52–1881; p = 0.016) compared to their ctDNA-negative counterparts (Fig. 2C). (A) Kaplan–Meier estimates of patients with stage II–III rectal cancer representing disease-free survival stratified by ctDNA status (negative or positive) post-NAT at week 11. (B) Kaplan–Meier estimates of disease-free survival stratified by ctDNA status (positive or negative) and radiological response post-NAT, prior to surgery (left panel). Correlation of ctDNA clearance post-NAT with radiological relapse (right panel). (C) Kaplan–Meier estimates of disease-free survival stratified by ctDNA status post-NAT and rNAR score. (D) Kaplan–Meier estimates of disease-free survival stratified by ctDNA status post-NAT and pCR status. The HRs and 95% CIs were calculated using the Cox proportional (Fig. 2A-B-C) and Cox Regression with Firth’s Penalized Likelihood (D) hazard model. The P value was obtained with Fisher’s exact test whereas the OR and 95% CIs with simple logistic regression. ctDNA circulating tumor DNA, Cl confidence interval, DFS disease-free survival, HR hazard ratio, NAT neoadjuvant therapy, pCR pathologic complete response, + positive, − negative. ctDNA detection rates at pre- and post-treatment time points, pathological parameters and association with patients outcomes ctDNA status at the post-NAT time point was noted to correlate with pathologic response. Patients who remained ctDNA-positive post-NAT and were unable to achieve pCR exhibited (N = 5) significantly inferior DFS (HR: 21.18, 95% CI 2.32–2804, p = 0.0037) and showed a 36-month DFS of 20% (95% CI 0.8–58.2%) compared to patients who remained ctDNA-negative, irrespective of pCR status (N = 25) and showed a 36-month DFS of 100% (ctDNA-negative and pCR N = 6) and 89.5% (95% CI 64.1–97.3%, ctDNA-negative and no pCR, N = 19) (Fig. 2D). As aforementioned, the NAR score has been shown to be a reliable measure of treatment response in rectal cancer after surgery. In our study, patients with a high NAR score showed a higher recurrence rate of 89% (8/9) compared to 33.3% (4/12) in the intermediate group, and 0% (0/9) in the low-risk group. Patients with a high NAR score exhibited a significantly inferior DFS when compared to those with low-risk (HR: 29.6, 95% CI 3.7–3838, p < 0.001) while the difference was statistically insignificant when comparing the low and intermediate-risk groups (HR: 7.33, 95% CI 0.78–972, p = 0.088) (Fig. 3A). (A) Kaplan-Meier estimates of patients with stage II–III rectal cancer representing NAR scores stratified by DFS. (B) Association between post-NAT ctDNA status and NAR score (i) and rates of recurrence based on patient NAR scores (low, intermediate, or high) and ctDNA status post-NAT (ii). The p-value for the association between post-NAT ctDNA and NAR score was assessed with the Wilcoxon rank sum two-sided test (left panel) and Fisher’s exact test (right panel). C. Kaplan-Meier estimates of patients with stage II–III rectal cancer representing disease-free survival stratified by post-NAT NAR scores and ctDNA status. HRs and 95% CIs were calculated using Cox Regression with Firth’s Penalized Likelihood (A,C) hazard model. P values were calculated using the two-sided log-rank test. ctDNA circulating tumor DNA, Cl confidence interval, DFS disease-free survival, HR hazard ratio, NAR neoadjuvant rectal score, NAT neoadjuvant therapy, + positive, − negative. In the post-NAT setting, patients who remained ctDNA-positive were more likely to have a higher NAR score (p = 0.004) (Fig. 3B-i). Specifically, patients with a high NAR score were more likely to be ctDNA-positive (44% [4/9]) compared to those in the intermediate (8% [1/12]) and low-risk (0% [9/9]) groups (p = 0.044) (Fig. 3B-ii). When utilizing ctDNA status as an adjunct to the NAR score within the post-NAT setting, patients who were ctDNA-positive with an intermediate/ high NAR score were found to have a higher recurrence rate of 100% (5/5) compared to those who were ctDNA-negative with either an intermediate/ high NAR score (43.8% [7/16]) or low NAR score (0% [0/9]). As expected, patients who were ctDNA-positive with an intermediate/high NAR score exhibited a significantly inferior 36-month DFS of 20% (95% CI 0.8–58.2%, HR: 47.5, 95% CI 5.1–6338, p < 0.001) when compared to those who were ctDNA-negative with a low NAR score (36-month DFS: 100%) (Fig. 3C). ctDNA detection rates post-surgery and association with patient outcomes Post-surgery, the ctDNA positivity rates were 10% (3/30). Out of these, two patients had stage III disease and all of them were ctDNA positive at baseline. The presence of ctDNA postoperatively, within the MRD window (2–12 weeks post-surgery) was associated with a higher recurrence rate of 100% (3/3) when compared to an event rate of 27.3% (6/22) in ctDNA-negative patients. Moreover, ctDNA-positive patients exhibited a significantly inferior DFS (HR: 19.65, 95% CI 3.17–121.75, p = 0.001) with a median DFS of 8.37 months for ctDNA-positive patients compared to NR for ctDNA-negative patients (Fig. 4A). Kaplan-Meier estimates of patients with stage II–III rectal cancer representing disease-free survival stratified by (A) ctDNA status at the MRD time point (week 2–12 post-surgery). (B) ctDNA dynamics during the post-NAT to MRD time point. To account for immortal time bias, a landmark analysis was performed 12 weeks after surgery. (C) Multivariate and univariate analysis of prognostic factors and their association with DFS, as indicated by HR, analyzed across the cohort; HR hazard ratio. ctDNA circulating tumor DNA, Cl confidence interval, DFS disease-free survival, HR hazard ratio, MRD molecular residual disease, NAR neoadjuvant rectal score, NAT neoadjuvant therapy. Interestingly, when evaluating ctDNA dynamics during the post-NAT to MRD time point, we found that patients who remained persistently negative had superior DFS when compared to those who remained persistently positive (HR: 15.9, 95% CI 2.6–98, p = 0.003) or who converted negative (HR: 8.5, 95% CI 2–36.4, p = 0.004) (Fig. 4B). Finally, in multivariate analysis, ctDNA status post-surgery at the MRD time point (p = 0.048) and the NAR score (p = 0.025) were found to be the most significant risk factors associated with disease recurrence (Fig. 4C). Discussion ctDNA has evolved into a promising prognostic biomarker across several tumor types. In this study, we demonstrate that ctDNA detection post-NAT and post-surgery is highly prognostic of poor outcomes in rectal cancer. Specifically, we found that ctDNA positivity at the MRD time point using a personalized, tumor-informed ctDNA assay identified recurrence with a positive predictive value (PPV) and specificity of 100%. This suggests that ctDNA may enable the prediction of treatment response and long-term survival outcomes in patients with rectal cancer. The persistence of ctDNA following chemoradiation therapy (CRT) and after surgery in patients with rectal cancer and its association with disease recurrence and poor survival has been reported in the literature17,20,22,23. Most recently, in a multicenter, open-label, prospective phase II/III randomized trial, Liu and colleagues compared different approaches for MRD detection and found the tumor-informed approach to be the most predictive, demonstrating ctDNA-positivity post-NAT to be associated with an increased risk of recurrence (HR = 27.38; P < 0.0001)6. In multivariate analysis, in comparing post-NAT ctDNA fraction with traditional clinicopathologic risk factors (age, sex, and clinical risk), post-NAT ctDNA fraction assessed using a tumor-informed assay was found to be the strongest independent predictor of recurrence6. Similarly, in a prospective biomarker study, Tie et al. observed that patients who were ctDNA-positive either post-CRT or postoperatively exhibited significantly poorer 3-year RFS when compared to ctDNA-negative patients (post-CRT: 50% versus 85%, HR: 6.6, 95% CI 2.6–17, p < 0.001; post-operative: 33% versus 87%, HR: 13, 95% CI 5.5–31, p < 0.001)21. Notably, they found post-surgery ctDNA positivity to be predictive of disease recurrence regardless of adjuvant therapy administration (ACT: HR 10.0; P < 0.001; without ACT: HR 22.0; P < 0.001)21. Despite the small number of patients, the findings obtained from our study are consistent with the published literature, reiterating the negative prognostic value of ctDNA positivity post-NAT and post-surgery. Post-NAT cCR is often used as a composite surrogate of NAT response, in place of pCR, which is dependent on post-surgical pathological review2. However, several studies have reported low concordance between radiological staging (Magnetic Resonance Tumor Regression Grade, mrTRG) and pathologic assessment (pTRG), with patients often under-staged by MRI17,24,25. As such, in our study, we found that patients who were ctDNA-negative post-NAT, irrespective of radiological response, demonstrated superior outcomes when compared to patients who remained ctDNA-positive. Moreover, ctDNA status in combination with rNAR score seems to be prognostic of long-term outcomes. Thus, modalities with higher predictive performance, such as ctDNA assessment, are needed to appropriately risk stratify patients post-NAT. With regard to the pathological response to NAT, several studies did not find any association between ctDNA status and pCR20,21,26,27, even if in many studies preoperative ctDNA presence are related to important pathological parameters, such as tumor burden22, surgical margins, lymph-node status and NAR score26. In the current study, we found that post-NAT ctDNA status correlated with pCR, as none of the patients who remained ctDNA-positive post-NAT achieved pCR. Furthermore, ctDNA-positive patients exhibited significantly inferior DFS (HR: 21.18, 95% CI 2.32–2804, p = 0.0037) compared to patients who remained ctDNA-negative, irrespective of pCR status (HR: 2.25, 95% CI 0.22–303, p = 0.554). Our findings show that ctDNA post-NAT can be used, at least, in conjunction with currently established clinical and pathologic response criteria to improve patient management, including Non Operative Management (NOM), as post-NAT ctDNA-positive patients should not be considered for NOM. Moreover, we independently validated the NAR score within our cohort and used it in combination with ctDNA status to predict treatment response and DFS. In keeping with published data11,28 we found the NAR score to be a marker of neoadjuvant treatment response and predictive of DFS. The addition of ctDNA status to this risk stratification algorithm broadened its application to include the post-surgical setting. Accordingly, we found that patients who were ctDNA-positive with intermediate to high NAR scores had higher rates of recurrence and poorer DFS when compared to those who were ctDNA-negative with intermediate to high or low NAR scores. Thus, we contend that post-treatment ctDNA status either as a sole outcome measure or as an adjunct to the NAR score has the potential to improve patient outcomes in rectal cancer. Although our data are largely applicable to clinical practice, there exist certain limitations including sample size, limited follow-up, and the potential for sub-optimal generalization of the study results due to analysis of only patients with tissue samples suitable for a tumor-informed ctDNA measurement. Regardless, we emphasize the prognostic and predictive value of personalized, tumor-informed ctDNA analysis in patients with rectal cancer and propose that ctDNA status post-NAT and post-surgery is a reliable predictor of treatment response and disease recurrence. We encourage clinical trials in patients with rectal cancer to incorporate tumor-informed ctDNA-based MRD assessment for early intervention to yield improved survival outcomes. Materials and methods Patients and samples A case series of 33 patients with stage II–III rectal cancer enrolled from 2017 to 2021 at Istituto Romagnolo per lo Studio dei Tumori “Dino Amadori” IRST-IRCCS were analyzed retrospectively. All patients received neoadjuvant chemoradiotherapy followed by surgery. Chemo-naive formalin-fixed paraffin-embedded (FFPE) tissues, peripheral blood and plasma samples obtained pre-NAT, post-NAT at and at end-of-treatment were collected, as well as clinicopathologic information and follow-up data for all patients. The study was performed in accordance with the Declaration of Helsinki and the Local Ethical Committees of Romagna (CE.ROM.) approved this study under reference number IRSTB081. All patients signed written informed consent for study participation. MRD assessment using a multiplex (m) PCR-NGS assay A clinically validated, personalized, tumor-informed, 16-plex PCR assay (Signatera™, Natera, Inc.) was used for the detection and quantification of ctDNA, as previously published13. Briefly, whole-exome sequencing (WES) was performed on FFPE tissue and matched normal DNA obtained from blood samples. Based on the results of WES, a patient-specific assay (targeting 16 somatic single-nucleotide variants) was used to track ctDNA in the associated patients’ plasma samples using an mPCR-NGS approach. Cell-free DNA was extracted from patient plasma at a given time point and was used to detect ctDNA. Plasma samples with at least two variants above the predefined detection threshold were deemed ctDNA-positive. The pre-defined threshold is based on Natera’s proprietary variant calling method wherein detecting at least two out of 16 variants ensures the optimal analytical performance of the assay. Statistical analysis The primary objective of the study was to correlate ctDNA status post-NAT (week 11) and post-surgery with disease-free survival (DFS). Our secondary objective included correlating ctDNA status post-NAT (week 11) and post-surgery with radiological response and rNAR score, as well as pathological parameters (pCR and NAR score). The primary outcome measure was DFS, measured from the date of surgery to the first documented sign of radiological disease progression, either locoregional or distant, or death from any cause, whichever occurred first. Patient characteristics were summarized using the median and first- and third quartiles (IQ, IIIQ) for continuous variables and natural frequencies and percentages for categorical ones. Statistical significance was evaluated using Fisher’s exact test for categorical variables and Wilcoxon rank sum two-sided test for continuous variables. The Kaplan-Meier method was used for estimating survival distributions. Hazard ratios (HR), associated 95% confidence intervals (CI), and p-values were calculated using Cox regression proportional hazards or Cox Regression with Firth’s Penalized Likelihood hazard model. The log-rank test was used for comparing two or more survival curves with p ≤ 0.05 considered significant. For the MRD time point analysis, to account for immortal time bias, a landmark analysis was performed at 12 weeks after surgery. Survival analyses were carried out using R software v4.2.2 using packages survival, survminer and coxphf for Firth’s penalized maximum likelihood. Odds ratio was calculated with simple logistic regression using the stats package in R. The NAR score was calculated using the following Eq. (1): where, cT (clinical tumor stage) is an element of the set {1, 2, 3, 4}, pT (pathologic tumor stage) is in {0, 1, 2, 3, 4}, and pN (pathologic nodal stage) is in {0, 1, 2}. Based on the NAR score, patients were stratified into low (NAR < 8), intermediate (NAR = 8–16), and high (NAR > 16) risk groups. The rNAR is a modification of NAR score, and it is calculated using only pre- and post-treatment clinical stages, instead of pathological ones12. Acknowledgements This work was partly supported thanks to the contribution of Ricerca Corrente by the Italian Ministry of Health within the research line “Precision, gender and ethnicity-based medicine and geroscience: genetic-molecular mechanisms in the development, characterization and treatment of tumors. The authors acknowledge the resources (sample and data) and services provided by the ‘Centro Risorse Biologiche (CRB) IRST’, a research facility of Istituto Scientifico Romagnolo per la Cura dei Tumori ‘Dino Amadori’ (BBMRI-ERIC ID: IT_1611958555932851). Authors would like to acknowledge funding from Natera through an Innovation Award Grant to Chiara Molinari. Author contributions C. M. and A.P conceptualized the projectc; IGR, API, GF, GLF, MM, AR and AP performed data curation; GL, SS, ES and EP performed formal analysis; CM acquired funding; CM, GM, AP, AJ and MCL conducted the investigation; GL, ES and SS designed, performed, and evaluated the experiments; CM, AP and SS administered the project; CM, GM, MC, LS and GG found resources; GL and SS used software; GLF, PU, GM, AJ, MCL supervised; GL and SS validated the data; CM, GM, AP, GL, SS, GVG, MM prepared the figures; CM, GM, AP, GVG, PD and MM wrote the originale draft and all authors reviewed the manuscript. Data availability The data underlying this article will be shared on reasonable request to the corresponding author.All the scripts used for the analysis of this cohort have been deposited in the following github public depository (https://github.com/Natera-TMED/Molinari-et-al_-Rectal-Cohort.git). Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement The data underlying this article will be shared on reasonable request to the corresponding author.All the scripts used for the analysis of this cohort have been deposited in the following github public depository (https://github.com/Natera-TMED/Molinari-et-al_-Rectal-Cohort.git).