A concurrent, deep learning-based computer-aided detection system for prostate multiparametric MRI: a performance study involving experienced and less-experienced radiologists

Performance of a Region of Interest-based Algorithm in Diagnosing International Society of Urological Pathology Grade Group ≥2 Prostate Cancer on the MRI-FIRST Database-CAD-FIRST Study

BACKGROUND AND OBJECTIVE

Prostate multiparametric magnetic resonance imaging (MRI) shows high sensitivity for International Society of Urological Pathology grade group (GG) ≥2 cancers. Many artificial intelligence algorithms have shown promising results in diagnosing clinically significant prostate cancer on MRI. To assess a region-of-interest-based machine-learning algorithm aimed at characterising GG ≥2 prostate cancer on multiparametric MRI.

METHODS

The lesions targeted at biopsy in the MRI-FIRST dataset were retrospectively delineated and assessed using a previously developed algorithm. The Prostate Imaging-Reporting and Data System version 2 (PI-RADSv2) score assigned prospectively before biopsy and the algorithm score calculated retrospectively in the regions of interest were compared for diagnosing GG ≥2 cancer, using the areas under the curve (AUCs), and sensitivities and specificities calculated with predefined thresholds (PIRADSv2 scores ≥3 and ≥4; algorithm scores yielding 90% sensitivity in the training database). Ten predefined biopsy strategies were assessed retrospectively.

KEY FINDINGS AND LIMITATIONS

After excluding 19 patients, we analysed 232 patients imaged on 16 different scanners; 85 had GG ≥2 cancer at biopsy. At patient level, AUCs of the algorithm and PI-RADSv2 were 77% (95% confidence interval [CI]: 70-82) and 80% (CI: 74-85; p = 0.36), respectively. The algorithm's sensitivity and specificity were 86% (CI: 76-93) and 65% (CI: 54-73), respectively. PI-RADSv2 sensitivities and specificities were 95% (CI: 89-100) and 38% (CI: 26-47), and 89% (CI: 79-96) and 47% (CI: 35-57) for thresholds of ≥3 and ≥4, respectively. Using the PI-RADSv2 score to trigger a biopsy would have avoided 26-34% of biopsies while missing 5-11% of GG ≥2 cancers. Combining prostate-specific antigen density, the PI-RADSv2 and algorithm's scores would have avoided 44-47% of biopsies while missing 6-9% of GG ≥2 cancers. Limitations include the retrospective nature of the study and a lack of PI-RADS version 2.1 assessment.

CONCLUSIONS AND CLINICAL IMPLICATIONS

The algorithm provided robust results in the multicentre multiscanner MRI-FIRST database and could help select patients for biopsy.

PATIENT SUMMARY

An artificial intelligence-based algorithm aimed at diagnosing aggressive cancers on prostate magnetic resonance imaging showed results similar to expert human assessment in a prospectively acquired multicentre test database.

Emerging and anticipated innovations in prostate cancer MRI and their impact on patient care

Prostate cancer (PCa) remains the leading malignancy affecting men, with over 3 million men living with the disease in the US, and an estimated 288,000 new cases and almost 35,000 deaths in 2023 in the United States alone. Over the last few decades, imaging has been a cornerstone in PCa care, with a crucial role in the detection, staging, and assessment of PCa recurrence or by guiding diagnostic or therapeutic interventions. To improve diagnostic accuracy and outcomes in PCa care, remarkable advancements have been made to different imaging modalities in recent years. This paper focuses on reviewing the main innovations in the field of PCa magnetic resonance imaging, including MRI protocols, MRI-guided procedural interventions, artificial intelligence algorithms and positron emission tomography, which may impact PCa care in the future.

Recent trends in AI applications for pelvic MRI: a comprehensive review

Magnetic resonance imaging (MRI) is an essential tool for evaluating pelvic disorders affecting the prostate, bladder, uterus, ovaries, and/or rectum. Since the diagnostic pathway of pelvic MRI can involve various complex procedures depending on the affected organ, the Reporting and Data System (RADS) is used to standardize image acquisition and interpretation. Artificial intelligence (AI), which encompasses machine learning and deep learning algorithms, has been integrated into both pelvic MRI and the RADS, particularly for prostate MRI. This review outlines recent developments in the use of AI in various stages of the pelvic MRI diagnostic pathway, including image acquisition, image reconstruction, organ and lesion segmentation, lesion detection and classification, and risk stratification, with special emphasis on recent trends in multi-center studies, which can help to improve the generalizability of AI.

Update on PI-RADS Version 2.1 Diagnostic Performance Benchmarks for Prostate MRI: Systematic Review and Meta-Analysis

Background Prostate MRI for the detection of clinically significant prostate cancer (csPCa) is standardized by the Prostate Imaging Reporting and Data System (PI-RADS), currently in version 2.1. A systematic review and meta-analysis infrastructure with a 12-month update cycle was established to evaluate the diagnostic performance of PI-RADS over time. Purpose To provide estimates of diagnostic accuracy and cancer detection rates (CDRs) of PI-RADS version 2.1 categories for prostate MRI, which is required for further evidence-based patient management. Materials and Methods A systematic search of PubMed, Embase, Cochrane Library, and multiple trial registers (English-language studies published from March 1, 2019, to August 30, 2022) was performed. Studies that reported data on diagnostic accuracy or CDRs of PI-RADS version 2.1 with csPCa as the primary outcome were included. For the meta-analysis, pooled estimates for sensitivity, specificity, and CDRs were derived from extracted data at the lesion level and patient level. Sensitivity and specificity for PI-RADS greater than or equal to 3 and PI-RADS greater than or equal to 4 considered as test positive were investigated. In addition to individual PI-RADS categories 1-5, subgroup analyses of subcategories (ie, 2+1, 3+0) were performed. Results A total of 70 studies (11 686 lesions, 13 330 patients) were included. At the patient level, with PI-RADS greater than or equal to 3 considered positive, meta-analysis found a 96% summary sensitivity (95% CI: 95, 98) and 43% specificity (95% CI: 33, 54), with an area under the summary receiver operating characteristic (SROC) curve of 0.86 (95% CI: 0.75, 0.93). For PI-RADS greater than or equal to 4, meta-analysis found an 89% sensitivity (95% CI: 85, 92) and 66% specificity (95% CI: 58, 74), with an area under the SROC curve of 0.89 (95% CI: 0.85, 0.92). CDRs were as follows: PI-RADS 1, 6%; PI-RADS 2, 5%; PI-RADS 3, 19%; PI-RADS 4, 54%; and PI-RADS 5, 84%. The CDR was 12% (95% CI: 7, 19) for transition zone 2+1 lesions and 19% (95% CI: 12, 29) for 3+0 lesions (P = .12). Conclusion Estimates of diagnostic accuracy and CDRs for PI-RADS version 2.1 categories are provided for quality benchmarking and to guide further evidence-based patient management. © RSNA, 2024 Supplemental material is available for this article. See also the editorial by Tammisetti and Jacobs in this issue.

Brain tumor detection with integrating traditional and computational intelligence approaches across diverse imaging modalities - Challenges and future directions

Brain tumor segmentation and classification play a crucial role in the diagnosis and treatment planning of brain tumors. Accurate and efficient methods for identifying tumor regions and classifying different tumor types are essential for guiding medical interventions. This study comprehensively reviews brain tumor segmentation and classification techniques, exploring various approaches based on image processing, machine learning, and deep learning. Furthermore, our study aims to review existing methodologies, discuss their advantages and limitations, and highlight recent advancements in this field. The impact of existing segmentation and classification techniques for automated brain tumor detection is also critically examined using various open-source datasets of Magnetic Resonance Images (MRI) of different modalities. Moreover, our proposed study highlights the challenges related to segmentation and classification techniques and datasets having various MRI modalities to enable researchers to develop innovative and robust solutions for automated brain tumor detection. The results of this study contribute to the development of automated and robust solutions for analyzing brain tumors, ultimately aiding medical professionals in making informed decisions and providing better patient care.

A deep learning-based semiautomated workflow for triaging follow-up MR scans in treated nasopharyngeal carcinoma

It is imperative to optimally utilize virtues and obviate defects of fully automated analysis and expert knowledge in new paradigms of healthcare. We present a deep learning-based semiautomated workflow (RAINMAN) with 12,809 follow-up scans among 2,172 patients with treated nasopharyngeal carcinoma from three centers (ChiCTR.org.cn, Chi-CTR2200056595). A boost of diagnostic performance and reduced workload was observed in RAINMAN compared with the original manual interpretations (internal vs. external: sensitivity, 2.5% [p = 0.500] vs. 3.2% [p = 0.031]; specificity, 2.9% [p < 0.001] vs. 0.3% [p = 0.302]; workload reduction, 79.3% vs. 76.2%). The workflow also yielded a triaging performance of 83.6%, with increases of 1.5% in sensitivity (p = 1.000) and 0.6%-1.3% (all p < 0.05) in specificity compared to three radiologists in the reader study. The semiautomated workflow shows its unique superiority in reducing radiologist's workload by eliminating negative scans while retaining the diagnostic performance of radiologists.

Diagnosis of thyroid disease using deep convolutional neural network models applied to thyroid scintigraphy images: a multicenter study

Objectives

The aim of this study was to improve the diagnostic performance of nuclear medicine physicians using a deep convolutional neural network (DCNN) model and validate the results with two multicenter datasets for thyroid disease by analyzing clinical single-photon emission computed tomography (SPECT) image data.

Methods

In this multicenter retrospective study, 3194 SPECT thyroid images were collected for model training (n=2067), internal validation (n=514) and external validation (n=613). First, four pretrained DCNN models (AlexNet, ShuffleNetV2, MobileNetV3 and ResNet-34) for were tested multiple medical image classification of thyroid disease types (i.e., Graves' disease, subacute thyroiditis, thyroid tumor and normal thyroid). The best performing model was then subjected to fivefold cross-validation to further assess its performance, and the diagnostic performance of this model was compared with that of junior and senior nuclear medicine physicians. Finally, class-specific attentional regions were visualized with attention heatmaps using gradient-weighted class activation mapping.

Results

Each of the four pretrained neural networks attained an overall accuracy of more than 0.85 for the classification of SPECT thyroid images. The improved ResNet-34 model performed best, with an accuracy of 0.944. For the internal validation set, the ResNet-34 model showed higher accuracy (p < 0.001) when compared to that of the senior nuclear medicine physician, with an improvement of nearly 10%. Our model achieved an overall accuracy of 0.931 for the external dataset, a significantly higher accuracy than that of the senior physician (0.931 vs. 0.868, p < 0.001).

Conclusion

The DCNN-based model performed well in terms of diagnosing thyroid scintillation images. The DCNN model showed higher sensitivity and greater specificity in identifying Graves' disease, subacute thyroiditis, and thyroid tumors compared to those of nuclear medicine physicians, illustrating the feasibility of deep learning models to improve the diagnostic efficiency for assisting clinicians.

The Present and Future of Artificial Intelligence in Urological Cancer

Artificial intelligence has drawn more and more attention for both research and application in the field of medicine. It has considerable potential for urological cancer detection, therapy, and prognosis prediction due to its ability to choose features in data to complete a particular task autonomously. Although the clinical application of AI is still immature and faces drawbacks such as insufficient data and a lack of prospective clinical trials, AI will play an essential role in individualization and the whole management of cancers as research progresses. In this review, we summarize the applications and studies of AI in major urological cancers, including tumor diagnosis, treatment, and prognosis prediction. Moreover, we discuss the current challenges and future applications of AI.

PI-RADS: Where Next?

Prostate MRI plays an important role in the clinical management of localized prostate cancer, mainly assisting in biopsy decisions and guiding biopsy procedures. The Prostate Imaging Reporting and Data System (PI-RADS) has been available to radiologists since 2012, with the most up-to-date and actively used version being PI-RADS version 2.1. This review article discusses the current use of PI-RADS, including its limitations and controversies, and summarizes research that aims to improve future iterations of this system.

A multicenter study of artificial intelligence-aided software for detecting visible clinically significant prostate cancer on mpMRI

BACKGROUND

AI-based software may improve the performance of radiologists when detecting clinically significant prostate cancer (csPCa). This study aims to compare the performance of radiologists in detecting MRI-visible csPCa on MRI with and without AI-based software.

MATERIALS AND METHODS

In total, 480 multiparametric MRI (mpMRI) images were retrospectively collected from eleven different MR devices, with 349 csPCa lesions in 180 (37.5%) cases. The csPCa areas were annotated based on pathology. Sixteen radiologists from four hospitals participated in reading. Each radiologist was randomly assigned to 30 cases and diagnosed twice. Half cases were interpreted without AI, and the other half were interpreted with AI. After four weeks, the cases were read again in switched mode. The mean diagnostic performance was compared using sensitivity and specificity on lesion level and patient level. The median reading time and diagnostic confidence were assessed.

RESULTS

On lesion level, AI-aided improved the sensitivity from 40.1% to 59.0% (18.9% increased; 95% confidence interval (CI) [11.5, 26.1]; p < .001). On patient level, AI-aided improved the specificity from 57.7 to 71.7% (14.0% increase, 95% CI [6.4, 21.4]; p < .001) while preserving the sensitivity (88.3% vs. 93.9%, p = 0.06). AI-aided reduced the median reading time of one case by 56.3% from 423 to 185 s (238-s decrease, 95% CI [219, 260]; p < .001), and the median diagnostic confidence score was increased by 10.3% from 3.9 to 4.3 (0.4-score increase, 95% CI [0.3, 0.5]; p < .001).

CONCLUSIONS

AI software improves the performance of radiologists by reducing false positive detection of prostate cancer patients and also improving reading times and diagnostic confidence.

CLINICAL RELEVANCE STATEMENT

This study involves the process of data collection, randomization and crossover reading procedure.

Characterization of high-grade prostate cancer at multiparametric MRI: assessment of PI-RADS version 2.1 and version 2 descriptors across 21 readers with varying experience (MULTI study)

OBJECTIVE

To assess PI-RADSv2.1 and PI-RADSv2 descriptors across readers with varying experience.

METHODS

Twenty-one radiologists (7 experienced (≥ 5 years) seniors, 7 less experienced seniors and 7 juniors) assessed 240 'predefined' lesions from 159 pre-biopsy multiparametric prostate MRIs. They specified their location (peripheral, transition or central zone) and size, and scored them using PI-RADSv2.1 and PI-RADSv2 descriptors. They also described and scored 'additional' lesions if needed. Per-lesion analysis assessed the 'predefined' lesions, using targeted biopsy as reference; per-lobe analysis included 'predefined' and 'additional' lesions, using combined systematic and targeted biopsy as reference. Areas under the curve (AUCs) quantified the performance in diagnosing clinically significant cancer (csPCa; ISUP ≥ 2 cancer). Kappa coefficients (κ) or concordance correlation coefficients (CCC) assessed inter-reader agreement.

RESULTS

At per-lesion analysis, inter-reader agreement on location and size was moderate-to-good (κ = 0.60-0.73) and excellent (CCC ≥ 0.80), respectively. Agreement on PI-RADSv2.1 scoring was moderate (κ = 0.43-0.47) for seniors and fair (κ = 0.39) for juniors. Using PI-RADSv2.1, juniors obtained a significantly lower AUC (0.74; 95% confidence interval [95%CI]: 0.70-0.79) than experienced seniors (0.80; 95%CI 0.76-0.84; p = 0.008) but not than less experienced seniors (0.74; 95%CI 0.70-0.78; p = 0.75). As compared to PI-RADSv2, PI-RADSv2.1 downgraded 17 lesions/reader (interquartile range [IQR]: 6-29), of which 2 (IQR: 1-3) were csPCa; it upgraded 4 lesions/reader (IQR: 2-7), of which 1 (IQR: 0-2) was csPCa. Per-lobe analysis, which included 60 (IQR: 25-73) 'additional' lesions/reader, yielded similar results.

CONCLUSIONS

Experience significantly impacted lesion characterization using PI-RADSv2.1 descriptors. As compared to PI-RADSv2, PI-RADSv2.1 tended to downgrade non-csPCa lesions, but this effect was small and variable across readers.

Artificial intelligence algorithms aimed at characterizing or detecting prostate cancer on MRI: How accurate are they when tested on independent cohorts? – A systematic review

PURPOSE

The purpose of this study was to perform a systematic review of the literature on the diagnostic performance, in independent test cohorts, of artificial intelligence (AI)-based algorithms aimed at characterizing/detecting prostate cancer on magnetic resonance imaging (MRI).

MATERIALS AND METHODS

Medline, Embase and Web of Science were searched for studies published between January 2018 and September 2022, using a histological reference standard, and assessing prostate cancer characterization/detection by AI-based MRI algorithms in test cohorts composed of more than 40 patients and with at least one of the following independency criteria as compared to the training cohort: different institution, different population type, different MRI vendor, different magnetic field strength or strict temporal splitting.

RESULTS

Thirty-five studies were selected. The overall risk of bias was low. However, 23 studies did not use predefined diagnostic thresholds, which may have optimistically biased the results. Test cohorts fulfilled one to three of the five independency criteria. The diagnostic performance of the algorithms used as standalones was good, challenging that of human reading. In the 12 studies with predefined diagnostic thresholds, radiomics-based computer-aided diagnosis systems (assessing regions-of-interest drawn by the radiologist) tended to provide more robust results than deep learning-based computer-aided detection systems (providing probability maps). Two of the six studies comparing unassisted and assisted reading showed significant improvement due to the algorithm, mostly by reducing false positive findings.

CONCLUSION

Prostate MRI AI-based algorithms showed promising results, especially for the relatively simple task of characterizing predefined lesions. The best management of discrepancies between human reading and algorithm findings still needs to be defined.