Rectal cancer staging: focus on the prognostic significance of the findings described by high-resolution magnetic resonance imaging

Performance reporting design in artificial intelligence studies using image-based TNM staging and prognostic parameters in rectal cancer: a systematic review

PURPOSE

The integration of artificial intelligence (AI) and magnetic resonance imaging in rectal cancer has the potential to enhance diagnostic accuracy by identifying subtle patterns and aiding tumor delineation and lymph node assessment. According to our systematic review focusing on convolutional neural networks, AI-driven tumor staging and the prediction of treatment response facilitate tailored treat-ment strategies for patients with rectal cancer.

METHODS

This paper summarizes the current landscape of AI in the imaging field of rectal cancer, emphasizing the performance reporting design based on the quality of the dataset, model performance, and external validation.

RESULTS

AI-driven tumor segmentation has demonstrated promising results using various convolutional neural network models. AI-based predictions of staging and treatment response have exhibited potential as auxiliary tools for personalized treatment strategies. Some studies have indicated superior performance than conventional models in predicting microsatellite instability and KRAS status, offer-ing noninvasive and cost-effective alternatives for identifying genetic mutations.

CONCLUSION

Image-based AI studies for rectal can-cer have shown acceptable diagnostic performance but face several challenges, including limited dataset sizes with standardized data, the need for multicenter studies, and the absence of oncologic relevance and external validation for clinical implantation. Overcoming these pitfalls and hurdles is essential for the feasible integration of AI models in clinical settings for rectal cancer, warranting further research.

Artificial intelligence and radiomics in magnetic resonance imaging of rectal cancer: a review

Rectal cancer (RC) is one of the most common tumours worldwide in both males and females, with significant morbidity and mortality rates, and it accounts for approximately one-third of colorectal cancers (CRCs). Magnetic resonance imaging (MRI) has been demonstrated to be accurate in evaluating the tumour location and stage, mucin content, invasion depth, lymph node (LN) metastasis, extramural vascular invasion (EMVI), and involvement of the mesorectal fascia (MRF). However, these features alone remain insufficient to precisely guide treatment decisions. Therefore, new imaging biomarkers are necessary to define tumour characteristics for staging and restaging patients with RC. During the last decades, RC evaluation via MRI-based radiomics and artificial intelligence (AI) tools has been a research hotspot. The aim of this review was to summarise the achievement of MRI-based radiomics and AI for the evaluation of staging, response to therapy, genotyping, prediction of high-risk factors, and prognosis in the field of RC. Moreover, future challenges and limitations of these tools that need to be solved to favour the transition from academic research to the clinical setting will be discussed.

Real-time Tracking and Classification of Tumor and Nontumor Tissue in Upper Gastrointestinal Cancers Using Diffuse Reflectance Spectroscopy for Resection Margin Assessment

Importance

Cancers of the upper gastrointestinal tract remain a major contributor to the global cancer burden. The accurate mapping of tumor margins is of particular importance for curative cancer resection and improvement in overall survival. Current mapping techniques preclude a full resection margin assessment in real time.

Objective

To evaluate whether diffuse reflectance spectroscopy (DRS) on gastric and esophageal cancer specimens can differentiate tissue types and provide real-time feedback to the operator.

Design, Setting, and Participants

This was a prospective ex vivo validation study. Patients undergoing esophageal or gastric cancer resection were prospectively recruited into the study between July 2020 and July 2021 at Hammersmith Hospital in London, United Kingdom. Tissue specimens were included for patients undergoing elective surgery for either esophageal carcinoma (adenocarcinoma or squamous cell carcinoma) or gastric adenocarcinoma.

Exposures

A handheld DRS probe and tracking system was used on freshly resected ex vivo tissue to obtain spectral data. Binary classification, following histopathological validation, was performed using 4 supervised machine learning classifiers.

Main Outcomes and Measures

Data were divided into training and testing sets using a stratified 5-fold cross-validation method. Machine learning classifiers were evaluated in terms of sensitivity, specificity, overall accuracy, and the area under the curve.

Results

Of 34 included patients, 22 (65%) were male, and the median (range) age was 68 (35-89) years. A total of 14 097 mean spectra for normal and cancerous tissue were collected. For normal vs cancer tissue, the machine learning classifier achieved a mean (SD) overall diagnostic accuracy of 93.86% (0.66) for stomach tissue and 96.22% (0.50) for esophageal tissue and achieved a mean (SD) sensitivity and specificity of 91.31% (1.5) and 95.13% (0.8), respectively, for stomach tissue and of 94.60% (0.9) and 97.28% (0.6) for esophagus tissue. Real-time tissue tracking and classification was achieved and presented live on screen.

Conclusions and Relevance

This study provides ex vivo validation of the DRS technology for real-time differentiation of gastric and esophageal cancer from healthy tissue using machine learning with high accuracy. As such, it is a step toward the development of a real-time in vivo tumor mapping tool for esophageal and gastric cancers that can aid decision-making of resection margins intraoperatively.

Preoperative Prediction of Perineural Invasion Status of Rectal Cancer Based on Radiomics Nomogram of Multiparametric Magnetic Resonance Imaging

Objectives

To compare the predictive performance of different radiomics signatures from multiparametric magnetic resonance imaging (mpMRI), including four sequences when used individually or combined, and to establish and validate an optimal nomogram for predicting perineural invasion (PNI) in rectal cancer (RC) patients.

Methods

Our retrospective study included 279 RC patients without preoperative antitumor therapy (194 in the training dataset and 85 in the test dataset) who underwent preoperative mpMRI scan between January 2017 and January 2021. Among them, 72 cases were PNI-positive. Then, clinical and radiological variables were collected, including carcinoembryonic antigen (CEA), radiological tumour stage (T_1-4), lymph node stage (N_0-2) and so on. Quantitative radiomics features were extracted and selected from oblique axial T₂-weighted imaging (T₂WI), T₁-weighted imaging (T₁WI), apparent diffusion coefficient (ADC), and enhanced T₁WI (T₁CE) sequences. The clinical model was constructed by integrating the final selected clinical and radiological variables. The radiomics signatures included four single-sequence signatures and one fusion signature were built using the respective remaining optimized features. And the nomogram was constructed based on the independent predictors by using multivariable logistic regression. The area under curve (AUC), DeLong test, calibration curve, and decision curve analysis (DCA) were used to evaluate the performance.

Results

Ultimately, 20 radiomics features were retained from the four sequences—T₁WI (n = 4), T₂WI (n = 5), ADC (n = 5), and T₁CE (n = 6)—to construct four single-sequence radiomics signatures and one fusion radiomics signature. The fusion radiomics signature performed better than four single-sequence radiomics signatures and clinical model (AUCs of 0.835 and 0.773 vs. 0.680-0.737 and 0.666-0.709 in the training and test datasets, respectively). The nomogram constructed by incorporating CEA, tumour stage and rad-score performed best, with AUCs of 0.869 and 0.864 in the training and test datasets, respectively. Delong test showed that the nomogram was significantly different from the clinical model and four single-sequence radiomics signatures (P < 0.05). Moreover, calibration curves demonstrated good agreement, and DCA highlighted benefits of the nomogram.

Conclusions

The comprehensive nomogram can preoperatively and noninvasively predict PNI status, provide a convenient and practical tool for treatment strategy, and help optimize individualized clinical decision-making in RC patients.

Magnetic resonance imaging-based artificial intelligence model in rectal cancer

Rectal magnetic resonance imaging (MRI) is the preferred method for the diagnosis of rectal cancer as recommended by the guidelines. Rectal MRI can accurately evaluate the tumor location, tumor stage, invasion depth, extramural vascular invasion, and circumferential resection margin. We summarize the progress of research on the use of artificial intelligence (AI) in rectal cancer in recent years. AI, represented by machine learning, is being increasingly used in the medical field. The application of AI models based on high-resolution MRI in rectal cancer has been increasingly reported. In addition to staging the diagnosis and localizing radiotherapy, an increasing number of studies have reported that AI models based on high-resolution MRI can be used to predict the response to chemotherapy and prognosis of patients.

Sex disparities in vitamin D status and the impact on systemic inflammation and survival in rectal cancer

BACKGROUND

We reported previously that rectal cancer patients given curative-intent chemotherapy, radiation, and surgery for non-metastatic disease had enhanced risk of metastatic progression and death if circulating levels of 25-hydroxyvitamin D [25(OH) D] were low. Here we investigated whether the association between the vitamin D status and prognosis pertains to the general, unselected population of rectal cancer patients.

METHODS

Serum 25(OH) D at the time of diagnosis was assessed in 129 patients, enrolled 2013-2017 and representing the entire range of rectal cancer stages, and analyzed with respect to season, sex, systemic inflammation, and survival.

RESULTS

In the population-based cohort residing at latitude 60°N, 25(OH) D varied according to season in men only, who were overrepresented among the vitamin D-deficient (< 50 nmol/L) patients. Consistent with our previous findings, the individuals presenting with T4 disease had significantly reduced 25(OH) D levels. Low vitamin D was associated with systemic inflammation, albeit with distinct modes of presentation. While men with low vitamin D showed circulating markers typical for the systemic inflammatory response (e.g., elevated erythrocyte sedimentation rate), the corresponding female patients had elevated serum levels of interleukin-6 and the chemokine (C-X-C motif) ligand 7. Despite disparities in vitamin D status and the potential effects on disease attributes, significantly shortened cancer-specific survival was observed in vitamin D-deficient patients irrespective of sex.

CONCLUSION

This unselected rectal cancer cohort confirmed the interconnection of low vitamin D, more advanced disease presentation, and poor survival, and further suggested it may be conditional on disparate modes of adverse systemic inflammation in men and women.

TRIAL REGISTRATION

ClinicalTrials.gov NCT01816607 ; registration date: 22 March 2013.

Preoperative staging of rectal cancer with MRI: correlation with pathologic staging

Introduction An accurate preoperative rectal cancer staging is crucial to the correct management of the disease. Despite great controversy around this issue, pelvic magnetic resonance (RM) is said to be the imagiologic standard modality. This work aimed to evaluate magnetic resonance accuracy in preoperative rectal cancer staging comparing with the anatomopathological results.

Methods We calculated sensibility, specificity, positive (VP positive) and negative (VP negative) predictive values for each T and N. We evaluated the concordance between both methods of staging using the Cohen weighted K (K w), and through ROC curves, we evaluated magnetic resonance accuracy in rectal cancer staging.

Results 41 patients met the inclusion criteria. We achieved an efficacy of 43.9% for T and 61% for N staging. The respective sensibility, specificity, positive and negative predictive values are 33.3%, 94.7%, 33.3% and 94.7% for T1; 62.5%, 32%, 37.0% and 57.1% for T2; 31.8%, 79%, 63.6% and 50% for T3 and 27.8%, 87%, 62.5% and 60.6% for N. We obtained a poor concordance for T and N staging and the anatomopathological results. The ROC curves indicated that magnetic resonance is ineffective in rectal cancer staging.

Conclusion Magnetic resonance has a moderate efficacy in rectal cancer staging and the major difficulty is in differentiating T2 and T3.

Accuracy of Magnetic Resonance Imaging in Staging Rectal Cancer with Multidisciplinary Team: A Single-Center Experience

Purpose: To investigate the accuracy of magnetic resonance imaging (MRI) in preoperative staging diagnosis for rectal cancer with multidisciplinary team (MDT) discussion. Methods: The retrospective study included 377 patients of rectal cancer with preoperative MRI staging from February 2015 to April 2018, in which 137 patients (36 received MDT discussion) received neoadjuvant therapy, 240 did not (97 received MDT discussion) and direct surgery was given. With postoperative pathological stage as the standard, the accuracy of MRI in preoperative staging for rectal cancer with MDT discussion was compared with non-MDT. Results: For direct surgery group, 21 out 97 (21.6%) patients changed their therapy strategy due to the change of the stage assessment after MDT. The accuracy of MRI for the diagnosis of preoperative N stage with MDT was significantly higher than those without MDT (56.2% vs. 42.1%, P=0.021). And for those without lymph node metastasis, the accuracy of MRI was higher after MDT (61.2% vs. 37.8%, P=0.009). For neoadjuvant therapy group, 7 out of 36 (19.4%) patients altered their therapy after MDT because of the changed stage. MDT improved the accuracy of restaging N stage with MRI (70.0% vs. 33.3%, P=0.003). The accuracy of MRI in staging T stage seemed not improved after MDT in both groups. Conclusions: In conclusion, MDT discussion increased the accuracy of MRI in preoperative staging diagnosis for rectal cancer. This mode could give a more accurate clinical stage of patients, which was in favor of choosing a preferable therapy strategy.

Differences in prognostic relevance of rectal magnetic resonance imaging findings before and after neoadjuvant chemoradiotherapy

This retrospective study was designed to compare prognostic relevance of magnetic resonance imaging (MRI) findings before and after neoadjuvant chemoradiotherapy (CRT). From 2002 to 2010, 399 patients who underwent surgery after CRT for rectal cancer (≥T3) and had adequate pre-CRT (mr) and post-CRT (ymr) MRI findings were examined. Factors examined included tumour (T), lymph node (N), mesorectal fascia (MRF), extramural venous invasion (EMVI), and tumour regression grade (TRG). Two Cox proportional hazard models were created using mr and ymr findings separately for overall survival (OS), disease-free survival (DFS), and local recurrence rate (LRR). Among mr findings, only mrEMVI was a significant prognostic factor for OS and DFS. Among ymr findings, ymrN, ymrMRF, and ymrEMVI were significant prognostic factors for OS and DFS, whereas ymrMRF and ymrEMVI were significant prognostic factors for LRR. C-indices tended to be higher for ymr findings than for mr findings (OS, 0.682 vs. 0.635; DFS, 0.660 vs. 0.631; LRR, 0.701 vs. 0.617). Survival outcomes of patients having all ymr risk factors were significantly poor (5-year OS, 52.4%; 5-year DFS, 38.1%; 5-year LRR, 27.7%). ymr findings showed better prognostic significance than mr findings. Among ymr findings, ymrN, ymrMRF, and ymrEMVI were independent prognostic factors for oncologic outcomes.

Determining the use of preoperative (chemo)radiotherapy in primary rectal cancer according to national and international guidelines

BACKGROUND

Pre-operative radiotherapy (RT) or chemoradiotherapy (CRT) is frequently used prior to rectal cancer surgery to improve local control and survival. The treatment is administered according to guidelines, but these recommendations vary significantly between countries. Based on the stage distribution and risk factors of rectal cancers as determined by magnetic resonance imaging (MRI) in an unselected Swedish population, the use of RT/CRT according to 15 selected guidelines is described.

MATERIALS AND METHODS

Selected guidelines from different countries and regions were applied to a well-characterized unselected population-based material of 686 primary non-metastatic rectal cancers staged by MRI. The fraction of patients assigned to surgery alone or surgery following pre-treatment with (C)RT was determined according to the respective guideline. RT/CRT administered to rectal cancer patients for other reasons, for example, for organ preservation or palliation, was not considered.

RESULTS

The fraction of patients with a clear recommendation for pre-treatment with (C)RT varied between 38% and 77% according to the different guidelines. In most guidelines, CRT was recommended to all patients who were not operated directly, and, in others, short-course RT was also recommended to patients with intermediate risk tumours. If only non-resectable or difficult to resect tumours were recommended pre-treatment, as stated in many Japanese publications, 9% would receive CRT followed by a delay to surgery.

CONCLUSIONS

According to most guidelines, well over 50% of primary non-metastatic rectal cancer patients from a general population, in which screening for colorectal cancer is not practised, are recommended treatment with pre-operative/neo-adjuvant therapy.

Fat-suppressed gadolinium-enhanced isotropic high-resolution 3D-GRE-T1WI for predicting small node metastases in patients with rectal cancer

Background

To investigate the application value of fat-suppressed gadolinium-enhanced isotropic high-resolution 3D-GRE-T1WI in regional nodes with different short-axis diameter ranges in rectal cancer, especially in nodes ≤5 mm.

Methods

Patients with rectal adenocarcinoma confirmed by postoperative histopathology were included, and all the patients underwent preoperative 3.0 T rectal magnetic resonance imaging (MRI) and total mesorectal excision (TME) within 2 weeks after an MR scan. The harvested nodes from specimens were matched with nodes in the field of view (FOV) of images for a node-by-node evaluation. The maximum short-axis diameters of all the visible nodes in the FOV of images were measured by a radiologist; the morphological and enhancement characteristics of these nodes were also independently evaluated by two radiologists. The χ² test was used to evaluate differences in morphological and enhancement characteristics between benign and malignant nodes. The enhancement characteristics were further compared between benign and malignant nodes with different short-axis diameter ranges using the χ² test. Kappa statistics were used to describe interobserver agreement.

Results

A total of 441 nodes from 70 enrolled patients were included in the evaluation, of which 111 nodes were metastatic. Approximately 85.5 and 95.6% of benign nodes were found to have obvious enhancement and homogeneous or mild-heterogeneous enhancement, respectively, whereas approximately 89.2 and 85.1% of malignant nodes showed moderate or mild enhancement and obvious-heterogeneous or rim-like enhancement, respectively. The area under the receiver operating characteristic (ROC) curve (AUC) values of the enhancement degree for identifying the overall nodal status, nodes ≤5 mm and nodes > 5 mm and ≤ 10 mm were 0.887, 0.859 and 0.766 for radiologist 1 and 0.892, 0.823 and 0.774 for radiologist 2, respectively. The AUCs of enhancement homogeneity were 0.940, 0.928 and 0.864 for radiologist 1 and 0.944, 0.938 and 0.842 for radiologist 2, respectively. Nodal border and signal homogeneity were also of certain value in distinguishing metastatic nodes.

Conclusions

Enhancement characteristics based on fat-suppressed gadolinium-enhanced isotropic high-resolution 3D-GRE-T1WI were helpful for diagnosing metastatic nodes in rectal cancer and were a reliable indicator for nodes ≤5 mm.

Pathologic processing of the total mesorectal excision

Total mesorectal excision (TME) is the current optimal surgical treatment for patients with rectal carcinoma. A complete TME is related to lower local recurrence rates and increased patient survival. Many confounding factors in the patient's anatomy and prior therapy can make it difficult to obtain a perfect plane, and thus a complete TME. The resection specimen can be thoroughly evaluated, grossly and microscopically, to identify substandard surgical outcomes and increased risk of local recurrence. Complete and accurate data reporting is critical for patient care and helps surgeons improve their technique.