Deep Learning Features Can Improve Radiomics-Based Prostate Cancer Aggressiveness Prediction.

PURPOSE: Emerging evidence suggests that the use of artificial intelligence can assist in the timely detection and optimization of therapeutic approach in patients with prostate cancer. The conventional perspective on radiomics encompassing segmentation and the extraction of radiomic features considers it as an independent and sequential process. However, it is not necessary to adhere to this viewpoint. In this study, we show that besides generating masks from which radiomic features can be extracted, prostate segmentation and reconstruction models provide valuable information in their feature space, which can improve the quality of radiomic signatures models for disease aggressiveness classification. MATERIALS AND METHODS: We perform 2,244 experiments with deep learning features extracted from 13 different models trained using different anatomic zones and characterize how modeling decisions, such as deep feature aggregation and dimensionality reduction, affect performance. RESULTS: While models using deep features from full gland and radiomic features consistently lead to improved disease aggressiveness prediction performance, others are detrimental. Our results suggest that the use of deep features can be beneficial, but an appropriate and comprehensive assessment is necessary to ensure that their inclusion does not harm predictive performance. CONCLUSION: The study findings reveal that incorporating deep features derived from autoencoder models trained to reconstruct the full prostate gland (both zonal models show worse performance than radiomics only models), combined with radiomic features, often lead to a statistically significant increase in model performance for disease aggressiveness classification. Additionally, the results also demonstrate that the choice of feature selection is key to achieving good performance, with principal component analysis (PCA) and PCA + relief being the best approaches and that there is no clear difference between the three proposed latent representation extraction techniques.

Analysis of domain shift in whole prostate gland, zonal and lesions segmentation and detection, using multicentric retrospective data.

Despite being one of the most prevalent forms of cancer, prostate cancer (PCa) shows a significantly high survival rate, provided there is timely detection and treatment. Computational methods can help make this detection process considerably faster and more robust. However, some modern machine-learning approaches require accurate segmentation of the prostate gland and the index lesion. Since performing manual segmentations is a very time-consuming task, and highly prone to inter-observer variability, there is a need to develop robust semi-automatic segmentation models. In this work, we leverage the large and highly diverse ProstateNet dataset, which includes 638 whole gland and 461 lesion segmentation masks, from 3 different scanner manufacturers provided by 14 institutions, in addition to other 3 independent public datasets, to train accurate and robust segmentation models for the whole prostate gland, zones and lesions. We show that models trained on large amounts of diverse data are better at generalizing to data from other institutions and obtained with other manufacturers, outperforming models trained on single-institution single-manufacturer datasets in all segmentation tasks. Furthermore, we show that lesion segmentation models trained on ProstateNet can be reliably used as lesion detection models.

Development of End-to-End AI-Based MRI Image Analysis System for Predicting IDH Mutation Status of Patients with Gliomas: Multicentric Validation.

Radiogenomics has shown potential to predict genomic phenotypes from medical images. The development of models using standard-of-care pre-operative MRI images, as opposed to advanced MRI images, enables a broader reach of such models. In this work, a radiogenomics model for IDH mutation status prediction from standard-of-care MRIs in patients with glioma was developed and validated using multicentric data. A cohort of 142 (wild-type: 32.4%) patients with glioma retrieved from the TCIA/TCGA was used to train a logistic regression model to predict the IDH mutation status. The model was evaluated using retrospective data collected in two distinct hospitals, comprising 36 (wild-type: 63.9%) and 53 (wild-type: 75.5%) patients. Model development utilized ROC analysis. Model discrimination and calibration were used for validation. The model yielded an AUC of 0.741 vs. 0.716 vs. 0.938, a sensitivity of 0.784 vs. 0.739 vs. 0.875, and a specificity of 0.657 vs. 0.692 vs. 1.000 on the training, test cohort 1, and test cohort 2, respectively. The assessment of model fairness suggested an unbiased model for age and sex, and calibration tests showed a p < 0.05. These results indicate that the developed model allows the prediction of the IDH mutation status in gliomas using standard-of-care MRI images and does not appear to hold sex and age biases.

Imaging of pancreatic ductal adenocarcinoma - An update for all stages of patient management.

Background: Pancreatic ductal adenocarcinoma (PDAC) is a common and lethal cancer. From diagnosis to disease staging, response to neoadjuvant therapy assessment and patient surveillance after resection, imaging plays a central role, guiding the multidisciplinary team in decision-planning. Review aims and findings: This review discusses the most up-to-date imaging recommendations, typical and atypical findings, and issues related to each step of patient management. Example cases for each relevant condition are presented, and a structured report for disease staging is suggested. Conclusion: Despite current issues in PDAC imaging at different stages of patient management, the radiologist is essential in the multidisciplinary team, as the conveyor of relevant imaging findings crucial for patient care.

Diffusion-weighted imaging for determining response to neoadjuvant therapy in pancreatic cancer: a systematic review and meta-analysis.

OBJECTIVES: To determine the role of diffusion-weighted imaging (DWI) for predicting response to neoadjuvant therapy (NAT) in pancreatic cancer. MATERIALS AND METHODS: MEDLINE, EMBASE, and Cochrane Library databases were searched for studies evaluating the performance of apparent diffusion coefficient (ADC) to assess response to NAT. Data extracted included ADC pre- and post-NAT, for predicting response as defined by imaging, histopathology, or clinical reference standards. ADC values were compared with standardized mean differences. Risk of bias was assessed using the Quality Assessment of Diagnostic Studies (QUADAS-2). RESULTS: Of 337 studies, 7 were included in the analysis (161 patients). ADC values reported for the pre- and post-NAT assessments overlapped between responders and non-responders. One study reported inability of ADC increase after NAT for distinguishing responders and non-responders. A correlation with histopathological response was reported for pre- and post-NAT ADC in 4 studies. DWI's diagnostic performance was reported to be high in three studies, with a 91.6-100% sensitivity and 62.5-94.7% specificity. Finally, heterogeneity and high risk of bias were identified across studies, affecting the domains of patient selection, index test, reference standard, and flow and timing. CONCLUSION: DWI might be useful for determining response to NAT in pancreatic cancer. However, there are still too few studies on this matter, which are also heterogeneous and at high risk for bias. Further studies with standardized procedures for data acquisition and accurate reference standards are needed. CLINICAL RELEVANCE STATEMENT: Diffusion-weighted MRI might be useful for assessing response to neoadjuvant therapy in pancreatic cancer. However, further studies with robust data are needed to provide specific recommendations for clinical practice. KEY POINTS: •The role of DWI with ADC measurements for assessing response to neoadjuvant therapy in pancreatic cancer is still unclear. •Pre- and post-neoadjuvant therapy ADC values overlap between responders and non-responders. •DWI has a reported high diagnostic performance for determining response when using histopathological or clinical reference standards; however, studies are still few and at high risk for bias.

Gastric schwannoma: Two case reports and review of the literature.

Gastric schwannomas are rare, slow-growing tumors whose clinical presentation is nonspecific. These are mostly benign, with a low probability of malignant transformation and an excellent prognosis. We present 2 cases of gastric schwannomas with distinct clinical features and imaging patterns, whose therapeutic approach differed. Case 1 is a 73-year-old woman with a voluminous subepithelial lesion in the greater gastric curvature, with predominantly endoluminal growth. Clinically the patient presented with nonspecific abdominal complaints and underwent complete surgical excision. Case 2 is a 69-year-old woman with an exophytic lesion adjacent to the gastric antrum, diagnosed incidentally and managed conservatively, with imaging follow-up, for the last 5 years and stable ever since. This article aims to focus on this rare disease, illustrating its main imaging findings, particularly in magnetic resonance imaging, along with pathological correlation, as well as reviewing the literature, discussing the differential diagnosis, and exploring clinical management and prognosis.

Urinary continence recovery after Retzius-sparing robot-assisted radical prostatectomy in relation to surgeon experience.

Urinary incontinence is one of the main concerns for patients after radical prostatectomy. Differences in surgical experience among surgeons could partly explain the wide range of frequencies observed. Our aim was to evaluate the association between the surgeons` experience and center caseload with relation to urinary continence recovery after Retzius-sparing robot-assisted radical prostatectomy (RS-RARP). Prospective observational single-center study. Five surgeons consecutively operated 405 patients between July 2017 and February 2022. Continence recovery was evaluated with pad count and by employing the short form of the International Consultation on Incontinence Questionnaire (ICIQ-SF), pre- and postoperatively at 1 year. Non-parametric tests were used. Median age was 63 years, 30% of patients presented with local advanced disease; the positive surgical margin rate (over 3 mm length) was 16%. Complication rate was 1% (Clavien-Dindo > II). One year after surgery, continence was assessed in 282 patients, of whom 87% were pad free and 51% never leaked (ICIQ-SF = 0). With respect to the mean annual number of procedures per surgeon, divided in < 20, 20-39 and ≥ 40, pad-free rates were achieved in 93%, 85%, and 84% and absence of urine leak rates in 47%, 62% and 48% of patients, respectively. Postoperative median ICIQ-SF was five. We acknowledge the limitation of a 12-month follow-up and the fact that we are a medium-volume center. There is no statistically significant association between continence recovery, surgeon's experience and center caseload. Continence recovery at 1 year after surgery is adequate and robust to surgeon's experience.

Performances and variability of CT radiomics for the prediction of microvascular invasion and survival in patients with HCC: a matter of chance or standardisation?

OBJECTIVES: To measure the performance and variability of a radiomics-based model for the prediction of microvascular invasion (MVI) and survival in patients with resected hepatocellular carcinoma (HCC), simulating its sequential development and application. METHODS: This study included 230 patients with 242 surgically resected HCCs who underwent preoperative CT, of which 73/230 (31.7%) were scanned in external centres. The study cohort was split into training set (158 patients, 165 HCCs) and held-out test set (72 patients, 77 HCCs), stratified by random partitioning, which was repeated 100 times, and by a temporal partitioning to simulate the sequential development and clinical use of the radiomics model. A machine learning model for the prediction of MVI was developed with least absolute shrinkage and selection operator (LASSO). The concordance index (C-index) was used to assess the value to predict the recurrence-free (RFS) and overall survivals (OS). RESULTS: In the 100-repetition random partitioning cohorts, the radiomics model demonstrated a mean AUC of 0.54 (range 0.44-0.68) for the prediction of MVI, mean C-index of 0.59 (range 0.44-0.73) for RFS, and 0.65 (range 0.46-0.86) for OS in the held-out test set. In the temporal partitioning cohort, the radiomics model yielded an AUC of 0.50 for the prediction of MVI, a C-index of 0.61 for RFS, and 0.61 for OS, in the held-out test set. CONCLUSIONS: The radiomics models had a poor performance for the prediction of MVI with a large variability in the model performance depending on the random partitioning. Radiomics models demonstrated good performance in the prediction of patient outcomes. CLINICAL RELEVANCE STATEMENT: Patient selection within the training set strongly influenced the performance of the radiomics models for predicting microvascular invasion; therefore, a random approach to partitioning a retrospective cohort into a training set and a held-out set seems inappropriate. KEY POINTS: • The performance of the radiomics models for the prediction of microvascular invasion and survival widely ranged (AUC range 0.44-0.68) in the randomly partitioned cohorts. • The radiomics model for the prediction of microvascular invasion was unsatisfying when trying to simulate its sequential development and clinical use in a temporal partitioned cohort imaged with a variety of CT scanners. • The performance of the radiomics models for the prediction of survival was good with similar performances in the 100-repetition random partitioning and temporal partitioning cohorts.

Value of handcrafted and deep radiomic features towards training robust machine learning classifiers for prediction of prostate cancer disease aggressiveness.

There is a growing piece of evidence that artificial intelligence may be helpful in the entire prostate cancer disease continuum. However, building machine learning algorithms robust to inter- and intra-radiologist segmentation variability is still a challenge. With this goal in mind, several model training approaches were compared: removing unstable features according to the intraclass correlation coefficient (ICC); training independently with features extracted from each radiologist's mask; training with the feature average between both radiologists; extracting radiomic features from the intersection or union of masks; and creating a heterogeneous dataset by randomly selecting one of the radiologists' masks for each patient. The classifier trained with this last resampled dataset presented with the lowest generalization error, suggesting that training with heterogeneous data leads to the development of the most robust classifiers. On the contrary, removing features with low ICC resulted in the highest generalization error. The selected radiomics dataset, with the randomly chosen radiologists, was concatenated with deep features extracted from neural networks trained to segment the whole prostate. This new hybrid dataset was then used to train a classifier. The results revealed that, even though the hybrid classifier was less overfitted than the one trained with deep features, it still was unable to outperform the radiomics model.

Retzius-sparing robot-assisted radical prostatectomy in a medium size oncological center holds adequate oncological and functional outcomes.

Retzius-sparing robot-assisted radical prostatectomy (RS-RARP) has emerged as a surgical option for patients with prostatic cancer in high-volume centers. The objective is to assess oncological and functional outcomes when implementing RS-RARP in a medium-volume center without previous experience of robotic surgery. This is a prospective observational single-center study. Patients operated between July 2017 and April 2020 were divided into two consecutive groups, A and B, each with 104 patients. The surgeons had prior experience in laparoscopic surgery and underwent robotic training. Positive surgical margin (PSM) status, urinary continence, and erectile function projected by Kaplan-Meier curves, together with patient reported quality of life outcomes at 12 months post-surgery were documented. Median patient age was 63 years (IQR = 59-67), overall PSM rate were 33%, 28% for pT2 disease. Pre-operative values showed no significant difference between both groups. The rate of urinary continence dropped from 81 to 78% (SE = 5.7) (Group A) and from 90 to 72% (SE = 6.3) (Group B) using the International Consultation on Incontinence Questionnaire-Short Form. Baseline sexual function was regained in 41% (Group A) and 47% (Group B) of patients. The median Expanded Prostate Index Composite-26 total score decreased from 86 to 82. These outcomes relate favorably to prior reports. There was a clinically significant decrease in median operative time in the successive groups with post-operative complications occurring in less than 2% of surgical procedures overall. A 12-month follow-up suggests that RS-RARP may be safely introduced in a medium-volume center without previous experience of robotic surgery.

Artificial intelligence and machine learning in cancer imaging.

An increasing array of tools is being developed using artificial intelligence (AI) and machine learning (ML) for cancer imaging. The development of an optimal tool requires multidisciplinary engagement to ensure that the appropriate use case is met, as well as to undertake robust development and testing prior to its adoption into healthcare systems. This multidisciplinary review highlights key developments in the field. We discuss the challenges and opportunities of AI and ML in cancer imaging; considerations for the development of algorithms into tools that can be widely used and disseminated; and the development of the ecosystem needed to promote growth of AI and ML in cancer imaging.

Challenging case of Whipple's disease: The contribution of radiology.

Whipple's disease is a rare chronic infectious disease, caused by Tropheryma whipplei. The disease can be challenging to diagnose due to the variable clinical manifestations and the nonspecific laboratory and imaging findings. We report the case of a 75-year-old man, complaining of weight loss and arthralgias with an insidious onset. A thoracic, abdominal and pelvic CT was performed, demonstrating features suggestive of Whipple's disease. Although not specific, the imaging findings of fatty attenuation mesenteric and retroperitoneal enlarged lymph nodes are a common finding in Whipple's disease. Esophagogastroduodenoscopy with duodenal biopsy confirmed the diagnosis through PCR analysis.

Prediction of Prostate Cancer Disease Aggressiveness Using Bi-Parametric Mri Radiomics.

Prostate cancer is one of the most prevalent cancers in the male population. Its diagnosis and classification rely on unspecific measures such as PSA levels and DRE, followed by biopsy, where an aggressiveness level is assigned in the form of Gleason Score. Efforts have been made in the past to use radiomics coupled with machine learning to predict prostate cancer aggressiveness from clinical images, showing promising results. Thus, the main goal of this work was to develop supervised machine learning models exploiting radiomic features extracted from bpMRI examinations, to predict biological aggressiveness; 288 classifiers were developed, corresponding to different combinations of pipeline aspects, namely, type of input data, sampling strategy, feature selection method and machine learning algorithm. On a cohort of 281 lesions from 183 patients, it was found that (1) radiomic features extracted from the lesion volume of interest were less stable to segmentation than the equivalent extraction from the whole gland volume of interest; and (2) radiomic features extracted from the whole gland volume of interest produced higher performance and less overfitted classifiers than radiomic features extracted from the lesions volumes of interest. This result suggests that the areas surrounding the tumour lesions offer relevant information regarding the Gleason Score that is ultimately attributed to that lesion.

Re-staging and follow-up of rectal cancer patients with MR imaging when "Watch-and-Wait" is an option: a practical guide.

In the past nearly 20 years, organ-sparing when no apparent viable tumour is present after neoadjuvant therapy has taken an increasingly relevant role in the therapeutic management of locally-advanced rectal cancer patients. The decision to include a patient or not in a "Watch-and-Wait" program relies mainly on endoscopic assessment by skilled surgeons, and MR imaging by experienced radiologists. Strict surveillance using the same modalities is required, given the chance of a local regrowth is of approximately 25-30%, almost always surgically salvageable if caught early. Local regrowths occur at the endoluminal aspect of the primary tumour bed in almost 90% of patients, but the rest are deep within it or outside the rectal wall, in which case detection relies solely on MR Imaging. In this educational review, we provide a practical guide for radiologists who are, or intend to be, involved in the re-staging and follow-up of rectal cancer patients in institutions with an established "Watch-and-Wait" program. First, we discuss patient preparation and MR imaging acquisition technique. Second, we focus on the re-staging MR imaging examination and review the imaging findings that allow us to assess response. Third, we focus on follow-up assessments of patients who defer surgery and confer about the early signs that may indicate a sustained/non-sustained complete response, a rectal/extra-rectal regrowth, and the particular prognosis of the "near-complete" responders. Finally, we discuss our proposed report template.

Two Rare Entities in One Patient: Mucinous Tubular and Spindle Cell Carcinoma of the Kidney and Peritoneal Adenomyomas.

Mucinous tubular and spindle cell carcinoma of the kidney is a rare subtype of renal cell carcinoma, that is believed to portend a favorable prognosis. Adenomyomas are benign tumors that typically arise from the myometrium. Extrauterine adenomyomas are extremely rare, with only a few cases reported in the literature. Here, we present an unusual case of a 46-year-old woman, with an incidentally detected bulky interpolar left kidney mass measuring 12 cm and multiple lobulated coalescent peritoneal nodules in the large epiploon suspicious for peritoneal carcinomatosis. A biopsy of the lesions revealed a mucinous tubular and spindle cell carcinoma of the kidney and extrauterine adenomyomas of the peritoneum. A left radical nephrectomy was performed and long-term hormone therapy with gonadotropin-releasing hormone agonists was prescribed. The purpose of this article is to focus on these two rare lesions, review the current literature, illustrate their key imaging findings along with pathologic correlation, as well as to discuss the differential diagnosis and clinical management.

Improving performance and generalizability in radiogenomics: a pilot study for prediction of IDH1/2 mutation status in gliomas with multicentric data.

Purpose: Radiogenomics offers a potential virtual and noninvasive biopsy. However, radiogenomics models often suffer from generalizability issues, which cause a performance degradation on unseen data. In MRI, differences in the sequence parameters, manufacturers, and scanners make this generalizability issue worse. Such image acquisition information may be used to define different environments and select robust and invariant radiomic features associated with the clinical outcome that should be included in radiomics/radiogenomics models. Approach: We assessed 77 low-grade gliomas and glioblastomas multiform patients publicly available in TCGA and TCIA. Radiomics features were extracted from multiparametric MRI images (T1-weighted, contrast-enhanced T1-weighted, T2-weighted, and fluid-attenuated inversion recovery) and different regions-of-interest (enhancing tumor, nonenhancing tumor/necrosis, and edema). A method developed to find variables that are part of causal structures was used for feature selection and compared with an embedded feature selection approach commonly used in radiomics/radiogenomics studies, across two different scenarios: (1) leaving data from a center as an independent held-out test set and tuning the model with the data from the remaining centers and (2) use stratified partitioning to obtain the training and the held-out test sets. Results: In scenario (1), the performance of the proposed methodology and the traditional embedded method was AUC: 0.75 [0.25; 1.00] versus 0.83 [0.50; 1.00], Sens.: 0.67 [0.20; 0.93] versus 0.67 [0.20; 0.93], Spec.: 0.75 [0.30; 0.95] versus 0.75 [0.30; 0.95], and MCC: 0.42 [0.19; 0.68] versus 0.42 [0.19; 0.68] for center 1 as the held-out test set. The performance of both methods for center 2 as the held-out test set was AUC: 0.64 [0.36; 0.91] versus 0.55 [0.27; 0.82], Sens.: 0.00 [0.00; 0.73] versus 0.00 [0.00; 0.73], Spec.: 0.82 [0.52; 0.94] versus 0.91 [0.62; 0.98], and MCC: - 0.13 [ - 0.38 ; - 0.04 ] versus - 0.09 [ - 0.38 ; - 0.02 ] , whereas for center 3 was AUC: 0.80 [0.62; 0.95] versus 0.89 [0.56; 0.96], Sens.: 0.86 [0.48; 0.97] versus 0.86 [0.48; 0.97], Spec.: 0.72 [0.54; 0.85] versus 0.79 [0.61; 0.90], and MCC: 0.47 [0.41; 0.53] versus 0.55 [0.48; 0.60]. For center 4, the performance of both methods was AUC: 0.77 [0.51; 1.00] versus 0.75 [0.47; 0.97], Sens.: 0.53 [0.30; 0.75] versus 0.00 [0.00; 0.15], Spec.: 0.71 [0.35; 0.91] versus 0.86 [0.48; 0.97], and MCC: 0.23 [0.16; 0.31] versus. - 0.32 [ - 0.46 ; - 0.20 ] . In scenario (2), the performance of these methods was AUC: 0.89 [0.71; 1.00] versus 0.79 [0.58; 0.94], Sens.: 0.86 [0.80; 0.92] versus 0.43 [0.15; 0.74], Spec.: 0.87 [0.62; 0.96] versus 0.87 [0.62; 0.96], and MCC: 0.70 [0.60; 0.77] versus 0.33 [0.24; 0.42]. Conclusions: This proof-of-concept study demonstrated good performance by the proposed feature selection method in the majority of the studied scenarios, as it promotes robustness of features included in the models and the models' generalizability by making used imaging data of different scanners or with sequence parameters.

Effective bowel motion reduction in mouse abdominal MRI using hyoscine butylbromide.

PURPOSE: Bowel motion is a significant source of artifacts in mouse abdominal MRI. Fasting and administration of hyoscine butylbromide (BUSC) have been proposed for bowel motion reduction but with inconsistent results and limited efficacy assessments. Here, we evaluate these regimes for mouse abdominal MRI at high field. METHODS: Thirty-two adult C57BL/6J mice were imaged on a 9.4T scanner with a FLASH sequence, acquired over 90 min with ~19 s temporal resolution. During MRI acquisition, 8 mice were injected with a low-dose and 8 mice with a high-dose bolus of BUSC (0.5 and 5 mg/kg, respectively). Eight mice were food deprived for 4.5-6.5 hours before MRI and another group of eight mice was injected with saline during MRI acquisition. Two expert readers reviewed the images and classified bowel motion, and quantitative voxel-wise analyses were performed for identification of moving regions. After defining the most effective protocol, high-resolution T2 -weighted and diffusion-weighted images were acquired from 4 mice. RESULTS: High-dose BUSC was the most effective protocol for bowel motion reduction, for up to 45 min. Fasting and saline protocols were not effective in suppressing bowel motion. High-resolution abdominal MRI clearly demonstrated improved image quality and ADC quantification with the high-dose BUSC protocol. CONCLUSION: Our data show that BUSC administration is advantageous for abdominal MRI in the mouse. Specifically, it endows significant bowel motion reduction, with relatively short onset timings after injection (~8.5 min) and relatively long duration of the effect (~45 min). These features improve the quality of high-resolution images of the mouse abdomen.

Early conformational changes at tumour bed and long term response after neoadjuvant therapy in locally-advanced rectal cancer.

OBJECTIVES: To evaluate how changes in tumour scar depth angle and thickness in the post-neoadjuvant period relate to long-term response in locally-advanced rectal cancer patients. METHODS: Informed consent was obtained from all patients and institutional review board approved this retrospective study. Sixty-nine consecutive locally-advanced rectal cancer patients who underwent neoadjuvant therapy and were selected for "Watch-and-Wait" were enrolled. Two radiologists, O1 and O2, blindly and independently reviewed the 1st and 2nd post-neoadjuvant therapy pelvic MRI T2-weighted images and recorded depth angle and thickness of the tumour scar. Value changes were calculated by simple subtraction (2nd-1st). Mann-Whitney U test was employed to assess for significant differences between sustained clinical complete responders (SCR), defined as patients with pathologic complete response or clinical complete response with a minimum follow-up of 1 year; and non-sustained complete responders (non-SCR). Interobserver agreement was estimated using intraclass correlation coefficient (ICC). Data on mrTRG, DWI and endoscopy at 1st and 2nd timepoints were retrieved for comparison. RESULTS: In SCR, depth angle change between 1st (med = 10 weeks after end of radiotherapy) and 2nd (med = 23 weeks after end of radiotherapy) timepoints was significantly different (O1:p = 0.004; O2:p = 0.010): the SCR group showed a depth angle reduction (O1:med=-4.45; O2:med=-2.35), whereas non-SCRs showed a depth angle increase (O1:med=+2.60; O2:med=+7.40). Also, at 2nd timepoint, SCR scars were significantly thinner both for O1 (p = 0.003; SCR:med = 7.05 mm; non-SCR:med = 9.4 mm) and O2 (p = 0.006; SCR:med = 6.45 mm; non-SCR:med = 8.2 mm). A depth angle increase >21º between 1st and 2nd timepoints and a scar thickness >10 mm at 2nd timepoint were not sensitive but were highly specific for a non-SCR (91/94 %) for both observers. Interobserver agreement was good for scar depth angle change (ICC = 0.65) and excellent for scar thickness at 2nd timepoint (ICC = 0.84). Of the retrieved data, only DWI at 2nd timepoint was discriminative (p = 0.043) providing a similar sensitivity (33 %) and a slightly lower specificity (87.5 %). CONCLUSION: Tumour scar expansion >21° between 1st and 2nd post-neoadjuvancy MRI and a scar thickness >10 mm at 2nd post-neoadjuvancy MRI may consistently indicate a non-SCR with high specificity in locally-advanced rectal cancer patients.