MRI radiomics in the prediction of therapeutic response to neoadjuvant therapy for locoregionally advanced rectal cancer: a systematic review

Development and validation of a competitive risk model in patients with rectal cancer: based on SEER database

BACKGROUND

Rectal cancer is one of the most common malignancies. To predict the specific mortality risk of rectal cancer patients, we constructed a predictive nomogram based on a competing risk model.

METHODS

The information on rectal cancer patients was extracted from the SEER database. Traditional survival analysis and specific death analysis were performed separately on the data.

RESULTS

The present study included 23,680 patients, with 16,580 in the training set and 7100 in the validation set. The specific mortality rate calculated by the competing risk model was lower than that of the traditional survival analysis. Age, Marriage, Race, Sex, ICD-O-3Hist/Behav, Grade, AJCC stage, T stage, N stage, Surgery, Examined LN, RX SUMM-SURG OTH, Chemotherapy, CEA, Deposits, Regional nodes positive, Brain, Bone, Liver, Lung, Tumor size, and Malignant were independent influencing factors of specific death. The overall C statistic of the model in the training set was 0.821 (Se = 0.001), and the areas under the ROC curve for cancer-specific survival (CSS) at 1, 3, and 5 years were 0.842, 0.830, and 0.812, respectively. The overall C statistic of the model in the validation set was 0.829 (Se = 0.002), and the areas under the ROC curve for CSS at 1, 3, and 5 years were 0.851, 0.836, and 0.813, respectively.

CONCLUSIONS

The predictive nomogram based on a competing risk model for time-specific mortality in patients with rectal cancer has very desirable accuracy. Thus, the application of the predictive nomogram in clinical practice can help physicians make clinical decisions and follow-up strategies.

MRI-Based Radiomic Models Outperform Radiologists in Predicting Pathological Complete Response to Neoadjuvant Chemoradiotherapy in Locally Advanced Rectal Cancer

RATIONALE AND OBJECTIVES

The 15%-27% of patients with locally advanced rectal cancer (LARC) achieved pathologic complete response (pCR) to neoadjuvant chemoradiotherapy (nCRT) and could avoid proctectomy. We aimed to investigate the effectiveness of treatment response prediction using MRI-based pre-, post-, and delta-radiomic features for LARC patients treated with nCRT and to compare these radiomic models with radiologists' visual assessment.

MATERIALS AND METHODS

A total of 126 patients with LARC who received nCRT before surgery were included and randomly divided into a training set (n = 84) and a validation set (n = 42). 250 radiomic features were extracted from T2-weighted images from pre- and post-nCRT MRI. Pearson correlation analysis and AONVA or Relief were used to identify radiomic descriptors associated with pCR. Five machine-learning classifiers were compared to construct radiomic models. The radiomic nomogram was built via multivariate logistic regression analysis. Two senior radiologists independently rated tumor regression grades and compared with radiomic models. Area under the curve (AUC) of the models and pooled observers were compared by using the DeLong test.

RESULTS

The optimal pre-, post-, and delta-radiomic models yielded an AUC of 0.717 (95% CI: 0.639-0.795), 0.805 (95%CI: 0.736-0.874), and 0.724 (95%CI: 0.648-0.800), respectively. The radiomic nomogram based on pre-nCRT cN stage, pre-nCRT radscore, and post-nCRT radscore achieved an AUC of 0.852 (95%CI: 0.774-0.930), which was higher than the single radiomic models and pooled readers (all p < 0.05).

CONCLUSIONS

The radiomic nomogram is an effective and invasive tool to predict pCR in LARC patients after nCRT, which outperforms radiologists.

Accuracy of deep neural learning models in the imaging prediction of pathological complete response after neoadjuvant chemoradiotherapy for locally advanced rectal cancer: a systematic review

PURPOSE

Up to 15-27% of patients achieve pathologic complete response (pCR) following neoadjuvant chemoradiotherapy (CRT) for locally advanced rectal cancer (LARC). Deep neural learning (DL) algorithms have been suggested to be a useful adjunct to allow accurate prediction of pCR and to identify patients who could potentially avoid surgery. This systematic review aims to interrogate the accuracy of DL algorithms at predicting pCR.

METHODS

Embase (PubMed, MEDLINE) databases and Google Scholar were searched to identify eligible English-language studies, with the search concluding in July 2022. Studies reporting on the accuracy of DL models in predicting pCR were selected for review and information pertaining to study characteristics and diagnostic measures was extracted from relevant studies. Risk of bias was evaluated using the Newcastle-Ottawa scale (NOS).

RESULTS

Our search yielded 85 potential publications. Nineteen full texts were reviewed, and a total of 12 articles were included in this systematic review. There were six retrospective and six prospective cohort studies. The most common DL algorithm used was the Convolutional Neural Network (CNN). Performance comparison was carried out via single modality comparison. The median performance for each best-performing algorithm was an AUC of 0.845 (range 0.71-0.99) and Accuracy of 0.85 (0.83-0.98).

CONCLUSIONS

There is a promising role for DL models in the prediction of pCR following neoadjuvant-CRT for LARC. Further studies are needed to provide a standardised comparison in order to allow for large-scale clinical application.

PROPERO REGISTRATION

PROSPERO 2021 CRD42021269904 Available from: https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42021269904 .

Validated Pretreatment Prediction Models for Response to Neoadjuvant Therapy in Patients with Rectal Cancer: A Systematic Review and Critical Appraisal

Pretreatment response prediction is crucial to select those patients with rectal cancer who will benefit from organ preservation strategies following (intensified) neoadjuvant therapy and to avoid unnecessary toxicity in those who will not. The combination of individual predictors in multivariable prediction models might improve predictive accuracy. The aim of this systematic review was to summarize and critically appraise validated pretreatment prediction models (other than radiomics-based models or image-based deep learning models) for response to neoadjuvant therapy in patients with rectal cancer and provide evidence-based recommendations for future research. MEDLINE via Ovid, Embase.com, and Scopus were searched for eligible studies published up to November 2022. A total of 5006 studies were screened and 16 were included for data extraction and risk of bias assessment using Prediction model Risk Of Bias Assessment Tool (PROBAST). All selected models were unique and grouped into five predictor categories: clinical, combined, genetics, metabolites, and pathology. Studies generally included patients with intermediate or advanced tumor stages who were treated with neoadjuvant chemoradiotherapy. Evaluated outcomes were pathological complete response and pathological tumor response. All studies were considered to have a high risk of bias and none of the models were externally validated in an independent study. Discriminative performances, estimated with the area under the curve (AUC), ranged per predictor category from 0.60 to 0.70 (clinical), 0.78 to 0.81 (combined), 0.66 to 0.91 (genetics), 0.54 to 0.80 (metabolites), and 0.71 to 0.91 (pathology). Model calibration outcomes were reported in five studies. Two collagen feature-based models showed the best predictive performance (AUCs 0.83-0.91 and good calibration). In conclusion, some pretreatment models for response prediction in rectal cancer show encouraging predictive potential but, given the high risk of bias in these studies, their value should be evaluated in future, well-designed studies.

Radiomic features derived from pretherapeutic MRI predict chemoradiation response in locally advanced rectal cancer

BACKGROUND

The standard treatment for locally advanced rectal cancer (LARC) is neoadjuvant concurrent chemoradiotherapy (CRT) followed by surgical excision. Current evidence suggests a favorable prognosis for those with pathological complete response (pCR), and surgery may be spared for them. We trained and validated regression models for CRT response prediction with selected radiomic features extracted from pretreatment magnetic resonance (MR) images to recruit potential candidates for this watch-and-wait strategy.

METHODS

We retrospectively enrolled patients with LARC who underwent pre-CRT MR imaging between 2010 and 2019. Pathological complete response in surgical specimens after CRT was defined as the ground truth. Quantitative features derived from both unfiltered and filtered images were extracted from manually segmented region of interests on T2-weighted images and selected using variance threshold, univariate statistical tests, and cross-validation least absolute shrinkage and selection operator (Lasso) regression. Finally, a regression model using selected features with high coefficients was optimized and evaluated. Model performance was measured by classification accuracies and area under the receiver operating characteristic (AUROC).

RESULTS

We extracted 1223 radiomic features from each MRI study of 133 enrolled patients. After tumor excision, 34 (26 %) of 133 patients had pCR in resected specimens. When 25 image-derived features were selected from univariate analysis, classification AUROC was 0.86 and 0.79 with the addition of six clinical features on the hold-out internal validation dataset. When 11 image-derived features were used, the optimized linear regression model had an AUROC value of 0.79 and 0.65 with the addition of six clinical features on the hold-out dataset. Among the radiomic features, texture features including gray level variance, strength, and cluster prominence had the highest coefficient by Lasso regression.

CONCLUSION

Radiomic features derived from pretreatment MR images demonstrated promising efficacy in predicting pCR after CRT. However, radiomic features combined with clinical features did not result in remarkable improvement in model performance.

Predicting Neoadjuvant Treatment Response in Rectal Cancer Using Machine Learning: Evaluation of MRI-Based Radiomic and Clinical Models

BACKGROUND

Radiomics is an approach to medical imaging that quantifies the features normally translated into visual display. While both radiomic and clinical markers have shown promise in predicting response to neoadjuvant chemoradiation therapy (nCRT) for rectal cancer, the interrelationship is not yet clear.

METHODS

A retrospective, single-institution study of patients treated with nCRT for locally advanced rectal cancer was performed. Clinical and radiomic features were extracted from electronic medical record and pre-treatment magnetic resonance imaging, respectively. Machine learning models were created and assessed for complete response and positive treatment effect using the area under the receiver operating curves.

RESULTS

Of 131 rectal cancer patients evaluated, 68 (51.9%) were identified to have a positive treatment effect and 35 (26.7%) had a complete response. On univariate analysis, clinical T-stage (OR 0.46, p = 0.02), lymphovascular/perineural invasion (OR 0.11, p = 0.03), and statin use (OR 2.45, p = 0.049) were associated with a complete response. Clinical T-stage (OR 0.37, p = 0.01), lymphovascular/perineural invasion (OR 0.16, p = 0.001), and abnormal carcinoembryonic antigen level (OR 0.28, p = 0.002) were significantly associated with a positive treatment effect. The clinical model was the strongest individual predictor of both positive treatment effect (AUC = 0.64) and complete response (AUC = 0.69). The predictive ability of a positive treatment effect increased by adding tumor and mesorectal radiomic features to the clinical model (AUC = 0.73).

CONCLUSIONS

The use of a combined model with both clinical and radiomic features resulted in the strongest predictive capability. With the eventual goal of tailoring treatment to the individual, both clinical and radiologic markers offer insight into identifying patients likely to respond favorably to nCRT.

Robustness of Radiomics in Pre-Surgical Computer Tomography of Non-Small-Cell Lung Cancer

Background: Radiomic features are increasingly used in CT of NSCLC. However, their robustness with respect to segmentation variability has not yet been demonstrated. The aim of this study was to assess radiomic features agreement across three kinds of segmentation. Methods: We retrospectively included 48 patients suffering from NSCLC who underwent pre-surgery CT. Two expert radiologists in consensus manually delineated three 3D-ROIs on each patient. To assess robustness for each feature, the intra-class correlation coefficient (ICC) across segmentations was evaluated. The ‘sensitivity’ of ICC upon some parameters affecting features computation (such as bin-width for first-order features and pixel-distances for second-order features) was also evaluated. Moreover, an assessment with respect to interpolator and isotropic resolution was also performed. Results: Our results indicate that ‘shape’ features tend to have excellent agreement (ICC > 0.9) across segmentations; moreover, they have approximately zero sensitivity to other parameters. ‘First-order’ features are in general sensitive to parameters variation; however, a few of them showed excellent agreement and low sensitivity (below 0.1) with respect to bin-width and pixel-distance. Similarly, a few second-order features showed excellent agreement and low sensitivity. Conclusions: Our results suggest that a limited number of radiomic features can achieve a high level of reproducibility in CT of NSCLC.

Combined artificial intelligence and radiologist model for predicting rectal cancer treatment response from magnetic resonance imaging: an external validation study

PURPOSE

To evaluate an MRI-based radiomic texture classifier alone and combined with radiologist qualitative assessment in predicting pathological complete response (pCR) using restaging MRI with internal training and external validation.

METHODS

Consecutive patients with locally advanced rectal cancer (LARC) who underwent neoadjuvant therapy followed by total mesorectal excision from March 2012 to February 2016 (Memorial Sloan Kettering Cancer Center/internal dataset, n = 114, 41% female, median age = 55) and July 2014 to October 2015 (Instituto do Câncer do Estado de São Paulo/external dataset, n = 50, 52% female, median age = 64.5) were retrospectively included. Two radiologists (R1, senior; R2, junior) independently evaluated restaging MRI, classifying patients (radiological complete response vs radiological partial response). Model A (n = 33 texture features), model B (n = 91 features including texture, shape, and edge features), and two combination models (model A + B + R1, model A + B + R2) were constructed. Pathology served as the reference standard for neoadjuvant treatment response. Comparison of the classifiers' AUCs on the external set was done using DeLong's test.

RESULTS

Models A and B had similar discriminative ability (P = 0.3; Model B AUC = 83%, 95% CI 70%-97%). Combined models increased inter-reader agreement compared with radiologist-only interpretation (κ = 0.82, 95% CI 0.70-0.89 vs k = 0.25, 95% CI 0.11-0.61). The combined model slightly increased junior radiologist specificity, positive predictive value, and negative predictive values (93% vs 90%, 57% vs 50%, and 91% vs 90%, respectively).

CONCLUSION

We developed and externally validated a combined model using radiomics and radiologist qualitative assessment, which improved inter-reader agreement and slightly increased the diagnostic performance of the junior radiologist in predicting pCR after neoadjuvant treatment in patients with LARC.

Dynamic Contrast-enhanced Magnetic Resonance Imaging Evaluation of Whole Tumour Perfusion Heterogeneity Predicts Distant Disease-free Survival in Locally Advanced Rectal Cancer

AIMS

To evaluate diffusion-weighted imaging and dynamic contrast-enhanced magnetic resonance imaging for the prediction of disease-free survival (DFS) in patients with locally advanced rectal cancer.

MATERIALS AND METHODS

Patients with stage II or III rectal adenocarcinoma undergoing neoadjuvant chemoradiotherapy (CRT) and surgery were eligible. Patients underwent multi-parametric magnetic resonance imaging (diffusion-weighted imaging and dynamic contrast-enhanced) before CRT, during CRT (week 3) and after CRT (1 week prior to surgery). Whole tumour apparent diffusion coefficient (ADC) and K^trans histogram quantiles (10th, 25th, 50th, 75th, 90th) were extracted for analysis. The associations between ADC and K^trans at three timepoints with time to relapse were analysed as a continuous variable using a Cox proportional hazard model.

RESULTS

Thirty-three patients were included in this analysis. The median follow-up was 4.4 years. No patient had locoregional relapse. Nine patients developed distant metastases. The hazard ratios for after CRT K^trans 10th (P = 0.035), 25th (P = 0.048), 50th (P = 0.046) and 75th (P = 0.045) quantiles were statistically significant for DFS. The best K^trans cut-off point after CRT for predicting relapse was 28 × 10^-3 mL/g/min (10th quantile), with a higher K^trans value predicting distant relapse. The 4-year DFS probability was 0.93 for patients with after CRT K^trans value ≤28 × 10^-3 mL/g/min versus 0.45 for patients with after CRT K^trans value >28 × 10^-3 mL/g/min. ADC was not able to predict DFS.

CONCLUSIONS

Patients with higher K^trans values after CRT (before surgery) in a histogram analysis of whole tumour heterogeneity had a significantly lower 4-year distant DFS and could be considered for more intense systemic therapy.

Simulation CT-based radiomics for prediction of response after neoadjuvant chemo-radiotherapy in patients with locally advanced rectal cancer

Background

To report on the discriminative ability of a simulation Computed Tomography (CT)-based radiomics signature for predicting response to treatment in patients undergoing neoadjuvant chemo-radiation for locally advanced adenocarcinoma of the rectum.

Methods

Consecutive patients treated at the Universities of Tübingen (from 1/1/07 to 31/12/10, explorative cohort) and Florence (from 1/1/11 to 31/12/17, external validation cohort) were considered in our dual-institution, retrospective analysis. Long-course neoadjuvant chemo-radiation was performed according to local policy. On simulation CT, the rectal Gross Tumor Volume was manually segmented. A feature selection process was performed yielding mineable data through an in-house developed software (written in Python 3.6). Model selection and hyper-parametrization of the model was performed using a fivefold cross validation approach. The main outcome measure of the study was the rate of pathologic good response, defined as the sum of Tumor regression grade (TRG) 3 and 4 according to Dworak’s classification.

Results

Two-hundred and one patients were included in our analysis, of whom 126 (62.7%) and 75 (37.3%) cases represented the explorative and external validation cohorts, respectively. Patient characteristics were well balanced between the two groups. A similar rate of good response to neoadjuvant treatment was obtained in in both cohorts (46% and 54.7%, respectively; p = 0.247). A total of 1150 features were extracted from the planning scans. A 5-metafeature complex consisting of Principal component analysis (PCA)-clusters (whose main components are LHL Grey-Level-Size-Zone: Large Zone Emphasis, Elongation, HHH Intensity Histogram Mean, HLL Run-Length: Run Level Variance and HHH Co-occurence: Cluster Tendency) in combination with 5-nearest neighbour model was the most robust signature. When applied to the explorative cohort, the prediction of good response corresponded to an average Area under the curve (AUC) value of 0.65 ± 0.02. When the model was tested on the external validation cohort, it ensured a similar accuracy, with a slightly lower predictive ability (AUC of 0.63).

Conclusions

Radiomics-based, data-mining from simulation CT scans was shown to be feasible and reproducible in two independent cohorts, yielding fair accuracy in the prediction of response to neoadjuvant chemo-radiation.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13014-022-02053-y.

The Utility of ADC First-Order Histogram Features for the Prediction of Metachronous Metastases in Rectal Cancer: A Preliminary Study

Simple Summary

Metachronous metastases are the main factors affecting survival in rectal cancer, and 15–25% of patients will develop them at a 5-year follow-up. Early identification of patients with higher risk of developing distant metachronous metastases would help to improve therapeutic protocols and could allow for a more accurate, personalized management. Apparent diffusion coefficient (ADC) represents an MRI quantitative biomarker, which can assess the diffusion characteristics of tissues, depending on the microscopic mobility of water, showing information related to tissue cellularity. First-order histogram-based features statistics describe the frequency distribution of intensity values within a region of interest, revealing microstructural alterations. In our study, we demonstrated that whole-tumor ADC first-order features may provide useful information for the assessment of rectal cancer prognosis, regarding the occurrence of metachronous metastases.

Abstract

This study aims the ability of first-order histogram-based features, derived from ADC maps, to predict the occurrence of metachronous metastases (MM) in rectal cancer. A total of 52 patients with pathologically confirmed rectal adenocarcinoma were retrospectively enrolled and divided into two groups: patients who developed metachronous metastases (n = 15) and patients without metachronous metastases (n = 37). We extracted 17 first-order (FO) histogram-based features from the pretreatment ADC maps. Student’s t-test and Mann–Whitney U test were used for the association between each FO feature and presence of MM. Statistically significant features were combined into a model, using the binary regression logistic method. The receiver operating curve analysis was used to determine the diagnostic performance of the individual parameters and combined model. There were significant differences in ADC 90th percentile, interquartile range, entropy, uniformity, variance, mean absolute deviation, and robust mean absolute deviation in patients with MM, as compared to those without MM (p values between 0.002–0.01). The best diagnostic was achieved by the 90th percentile and uniformity, yielding an AUC of 0.74 [95% CI: 0.60–0.8]). The combined model reached an AUC of 0.8 [95% CI: 0.66–0.90]. Our observations point out that ADC first-order features may be useful for predicting metachronous metastases in rectal cancer.

Sources of variation in multicenter rectal MRI data and their effect on radiomics feature reproducibility

OBJECTIVES

To investigate sources of variation in a multicenter rectal cancer MRI dataset focusing on hardware and image acquisition, segmentation methodology, and radiomics feature extraction software.

METHODS

T2W and DWI/ADC MRIs from 649 rectal cancer patients were retrospectively acquired in 9 centers. Fifty-two imaging features (14 first-order/6 shape/32 higher-order) were extracted from each scan using whole-volume (expert/non-expert) and single-slice segmentations using two different software packages (PyRadiomics/CapTk). Influence of hardware, acquisition, and patient-intrinsic factors (age/gender/cTN-stage) on ADC was assessed using linear regression. Feature reproducibility was assessed between segmentation methods and software packages using the intraclass correlation coefficient.

RESULTS

Image features differed significantly (p < 0.001) between centers with more substantial variations in ADC compared to T2W-MRI. In total, 64.3% of the variation in mean ADC was explained by differences in hardware and acquisition, compared to 0.4% by patient-intrinsic factors. Feature reproducibility between expert and non-expert segmentations was good to excellent (median ICC 0.89-0.90). Reproducibility for single-slice versus whole-volume segmentations was substantially poorer (median ICC 0.40-0.58). Between software packages, reproducibility was good to excellent (median ICC 0.99) for most features (first-order/shape/GLCM/GLRLM) but poor for higher-order (GLSZM/NGTDM) features (median ICC 0.00-0.41).

CONCLUSIONS

Significant variations are present in multicenter MRI data, particularly related to differences in hardware and acquisition, which will likely negatively influence subsequent analysis if not corrected for. Segmentation variations had a minor impact when using whole volume segmentations. Between software packages, higher-order features were less reproducible and caution is warranted when implementing these in prediction models.

KEY POINTS

• Features derived from T2W-MRI and in particular ADC differ significantly between centers when performing multicenter data analysis. • Variations in ADC are mainly (> 60%) caused by hardware and image acquisition differences and less so (< 1%) by patient- or tumor-intrinsic variations. • Features derived using different image segmentations (expert/non-expert) were reproducible, provided that whole-volume segmentations were used. When using different feature extraction software packages with similar settings, higher-order features were less reproducible.

Radiomics as a New Frontier of Imaging for Cancer Prognosis: A Narrative Review

The evaluation of the efficacy of different therapies is of paramount importance for the patients and the clinicians in oncology, and it is usually possible by performing imaging investigations that are interpreted, taking in consideration different response evaluation criteria. In the last decade, texture analysis (TA) has been developed in order to help the radiologist to quantify and identify parameters related to tumor heterogeneity, which cannot be appreciated by the naked eye, that can be correlated with different endpoints, including cancer prognosis. The aim of this work is to analyze the impact of texture in the prediction of response and in prognosis stratification in oncology, taking into consideration different pathologies (lung cancer, breast cancer, gastric cancer, hepatic cancer, rectal cancer). Key references were derived from a PubMed query. Hand searching and clinicaltrials.gov were also used. This paper contains a narrative report and a critical discussion of radiomics approaches related to cancer prognosis in different fields of diseases.