Diagnostic performance and reproducibility of T2w based and diffusion weighted imaging (DWI) based PI-RADSv2 lexicon descriptors for prostate MRI

Assessing the performance of ChatGPT and Bard/Gemini against radiologists for Prostate Imaging-Reporting and Data System classification based on prostate multiparametric MRI text reports

OBJECTIVES

Large language models (LLMs) have shown potential for clinical applications. This study assesses their ability to assign Prostate Imaging-Reporting and Data System (PI-RADS) categories based on clinical text reports.

METHODS

One hundred consecutive biopsy-naïve patients' multiparametric prostate MRI reports were independently classified by 2 uroradiologists, ChatGPT-3.5 (GPT-3.5), ChatGPT-4o mini (GPT-4), Bard, and Gemini. Original report classifications were considered definitive.

RESULTS

Out of 100 MRIs, 52 were originally reported as PI-RADS 1-2, 9 PI-RADS 3, 19 PI-RADS 4, and 20 PI-RADS 5. Radiologists demonstrated 95% and 90% accuracy, while GPT-3.5 and Bard both achieved 67%. Accuracy of the updated versions of LLMs increased to 83% (GTP-4) and 79% (Gemini), respectively. In low suspicion studies (PI-RADS 1-2), Bard and Gemini (F1: 0.94, 0.98, respectively) outperformed GPT-3.5 and GTP-4 (F1:0.77, 0.94, respectively), whereas for high probability MRIs (PI-RADS 4-5), GPT-3.5 and GTP-4 (F1: 0.95, 0.98, respectively) outperformed Bard and Gemini (F1: 0.71, 0.87, respectively). Bard assigned a non-existent PI-RADS 6 "hallucination" for 2 patients. Inter-reader agreements (Κ) between the original reports and the senior radiologist, junior radiologist, GPT-3.5, GTP-4, BARD, and Gemini were 0.93, 0.84, 0.65, 0.86, 0.57, and 0.81, respectively.

CONCLUSIONS

Radiologists demonstrated high accuracy in PI-RADS classification based on text reports, while GPT-3.5 and Bard exhibited poor performance. GTP-4 and Gemini demonstrated improved performance compared to their predecessors.

ADVANCES IN KNOWLEDGE

This study highlights the limitations of LLMs in accurately classifying PI-RADS categories from clinical text reports. While the performance of LLMs has improved with newer versions, caution is warranted before integrating such technologies into clinical practice.

Are T2WI PI-RADS sub-scores of transition zone prostate lesions biased by DWI information? A multi-reader, single-center study

PURPOSE

According to PI-RADS v2.1, T2-weighted imaging (T2WI) is the dominant sequence for transition zone (TZ) lesions. This study aimed to assess, whether diffusion-weighted imaging (DWI) information influences the assignment of T2WI scores.

METHOD

Out of 283 prostate MRI examinations with correlated biopsy results, fourty-four patients were selected retrospectively: first, 22 patients with a TZ lesion with T2WI and DWI scores ≥ 4, to represent lesions with unequivocal suspicion on T2WI and DWI. Second, 22 additional patients with TZ lesions of similar T2WI appearance but with corresponding DWI score ≤ 3 were added as control. Four residents and one board-certified radiologist each performed two assessments of the included patients: First, only T2WI was available (T2-only read); second, both T2WI and DWI sequences were available (biparametric read). Lesion scores were assessed using Wilcoxon signed-rank test, inter-reader agreement using weighted kappa and Kendall's W statistics, and sensitivity/specificity using McNemar test.

RESULTS

The T2WI scores were significantly different between the T2-only and biparametric read for 3 out of 4 residents (p ≤ 0.049) but not for the radiologist. The overall PI-RADS scores derived from the two reading sessions differed considerably for 35/220 cases (all readers pooled). Inter-reader agreement was fair for the T2WI and overall PI-RADS scores (mean kappa 0.27-0.30) and moderate for the DWI scores (mean kappa 0.43).

CONCLUSIONS

For inexperienced readers, assessment of T2WI is variable and potentially biased by availability of DWI information, which can lead to changes of overall PI-RADS score and consequently clinical management. Assessment of T2WI should be performed before reviewing DWI to ensure non-biased interpretation of TZ lesions in the dominant sequence.

Diagnostic Accuracy and Interobserver Agreement of PI-RADS Version 2 and Version 2.1 for the Detection of Transition Zone Prostate Cancers

BACKGROUND. PI-RADS version 2.1 (v2.1) introduced a number of key changes to the assessment of transition zone (TZ) lesions. OBJECTIVE. The purpose of this study was to evaluate interobserver agreement and diagnostic accuracy for detecting TZ prostate cancer (PCa) and clinically significant PCa (csPCa) by use of PI-RADS v2 and PI-RADS v2.1 among radiologists with different levels of experience. METHODS. This retrospective study included 355 biopsy-naïve patients who from January 2017 to March 2020 underwent prostate MRI that showed a TZ lesion and underwent subsequent biopsy. PCa was diagnosed in 93 patients (International Society of Urological Pathology [ISUP] grade group 1, n = 34; ISUP grade group ≥ 2, n = 59) and non-cancerous lesions in 262 patients. Five radiologists with varying experience in prostate MRI scored lesions using PI-RADS v2 and PI-RADS v2.1 in sessions separated by at least 4 weeks. Interobserver agreement was evaluated with kappa and Kendall W statistics. ROC curve analysis was used to evaluate performance in detection of TZ PCa and csPCa. RESULTS. Interobserver agreement among all readers was higher for PI-RADS v2.1 than for PI-RADS v2 (mean weighted κ = 0.700 vs 0.622; Kendall W = 0.805 vs 0.728; p = .03). The pooled AUC values for detecting TZ PCa and csPCa were higher among all readers using PI-RADS v2.1 (0.866 vs 0.827 for TZ PCa; 0.929 vs 0.899 for TZ csPCa; p < .001). For detecting TZ PCa, the pooled sensitivity, specificity, and accuracy were 86.9%, 79.4%, and 75.4% among all readers for PI-RADS v2.1 compared with 79.4%, 71.8%, and 73.8% for PI-RADS v2. For detecting TZ csPCa, the pooled sensitivity, specificity, and accuracy were 84.8%, 90.9%, and 89.9% among all readers for PI-RADS v2.1 compared with 81.4%, 89.9%, and 88.5% for PI-RADS v2. Reader 1, who had the least experience, had the lowest sensitivity, specificity, and accuracy (78.0%, 89.2%, and 87.3%). Reader 5, who had the most experience, had the highest sensitivity, specificity, and accuracy (88.1%, 92.9%, and 92.1%) in detecting csPCa. CONCLUSION. PI-RADS v2.1 had better interobserver agreement and diagnostic accuracy than PI-RADS v2 for evaluating TZ lesions. Reader experience continues to affect the performance of prostate MRI interpretation with PI-RADS v2.1. CLINICAL IMPACT. PI-RADS v2.1 is more accurate and reproducible than PI-RADS v2 for the diagnosis of TZ PCa.

Prostate Imaging-Reporting and Data System: Comparison of the Diagnostic Performance between Version 2.0 and 2.1 for Prostatic Peripheral Zone

Objective

To compare the diagnostic performance between Prostate Imaging-Reporting and Data System version 2.0 (PI-RADSv2.0) and version 2.1 (PI-RADSv2.1) for clinically significant prostate cancer (csPCa) in the peripheral zone (PZ).

Materials and Methods

This retrospective study included 317 patients who underwent multiparametric magnetic resonance imaging and targeted biopsy for PZ lesions. Definition of csPCa was International Society of Urologic Pathology grade ≥ 2 cancer. Area under the curve (AUC), sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy for csPCa were analyzed by two readers. The cancer detection rate (CDR) for csPCa was investigated according to the PI-RADS categories.

Results

AUC of PI-RADSv2.1 (0.856 and 0.858 for reader 1 and 2 respectively) was higher than that of PI-RADSv2.0 (0.795 and 0.747 for reader 1 and 2 respectively) (both p < 0.001). Sensitivity, specificity, PPV, NPV, and accuracy for PI-RADSv2.0 vs. PI-RADSv2.1 were 93.2% vs. 88.3% (p = 0.023), 52.8% vs. 76.6% (p < 0.001), 48.7% vs. 64.5% (p < 0.001), 94.2% vs. 93.2% (p = 0.504), and 65.9% vs. 80.4% (p < 0.001) for reader 1, and 96.1% vs. 92.2% (p = 0.046), 34.1% vs. 72.4% (p < 0.001), 41.3% vs. 61.7% (p < 0.001), 94.8% vs. 95.1% (p = 0.869), and 54.3% vs. 78.9% (p < 0.001) for reader 2, respectively. CDRs of PI-RADS categories 1–2, 3, 4, and 5 for PI-RADSv2.0 vs. PI-RADSv2.1 were 5.9% vs. 5.9%, 5.8% vs. 12.5%, 39.8% vs. 56.2%, and 88.9% vs. 88.9% for reader 1; and 4.5% vs. 4.1%, 6.1% vs. 11.1%, 32.5% vs. 53.4%, and 85.0% vs. 86.8% for reader 2, respectively.

Conclusion

Our data demonstrated improved AUC, specificity, PPV, accuracy, and CDRs of category 3 or 4 of PI-RADSv2.1, but decreased sensitivity, compared with PI-RADSv2.0, for csPCa in PZ.

Artificial Intelligence in Magnetic Resonance Imaging-based Prostate Cancer Diagnosis: Where Do We Stand in 2021?

CONTEXT

Men suspected of harboring prostate cancer (PCa) increasingly undergo multiparametric magnetic resonance imaging (mpMRI) and mpMRI-guided biopsy. The potential of mpMRI coupled to artificial intelligence (AI) methods to detect and classify PCa before decision-making requires investigation.

OBJECTIVE

To review the literature for studies addressing the diagnostic performance of combined mpMRI and AI approaches to detect and classify PCa, and to provide selection criteria for relevant articles having clinical significance.

EVIDENCE ACQUISITION

We performed a nonsystematic search of the English language literature using the PubMed-MEDLINE database up to October 30, 2020. We included all original studies addressing the diagnostic accuracy of mpMRI and AI to detect and classify PCa with histopathological analysis as a reference standard.

EVIDENCE SYNTHESIS

Eleven studies assessed AI and mpMRI approaches for PCa detection and classification based on a ground truth that referred to the entire prostate either with radical prostatectomy specimens (RPS) or relocalization of positive systematic and/or targeted biopsy. Seven studies retrospectively annotated cancerous lesions onto mpMRI identified in whole-mount sections from RPS, three studies used a backward projection of histological prostate biopsy information, and one study used a combined cohort of both approaches. All studies cross-validated their data sets; only four used a test set and one a multisite validation scheme. Performance metrics for lesion detection ranged from 87.9% to 92% at a threshold specificity of 50%. The lesion classification accuracy of the algorithms was comparable to that of the Prostate Imaging-Reporting and Data System.

CONCLUSIONS

For an algorithm to be implemented into radiological workflows and to be clinically applicable, it must be trained with a ground truth labeling that reflects histopathological information for the entire prostate and it must be externally validated. Lesion detection and classification performance metrics are promising but require prospective implementation and external validation for clinical significance.

PATIENT SUMMARY

We reviewed the literature for studies on prostate cancer detection and classification using magnetic resonance imaging (MRI) and artificial intelligence algorithms. The main application is in supporting radiologists in interpreting MRI scans and improving the diagnostic performance, so that fewer unnecessary biopsies are carried out.

PI-RADS Version 2.1: A Critical Review, From the AJR Special Series on Radiology Reporting and Data Systems

PI-RADS version 2.1 updates the technical parameters for multiparametric MRI (mpMRI) of the prostate and revises the imaging interpretation criteria while maintaining the framework introduced in version 2. These changes have been considered an improvement, although some issues remain unresolved, and new issues have emerged. Areas for improvement discussed in this review include the need for more detailed mpMRI protocols with optimization for 1.5-T and 3-T systems; lack of validation of revised transition zone interpretation criteria and need for clarifications of the revised DWI and dynamic contrast-enhanced imaging criteria and central zone (CZ) assessment; the need for systematic evaluation and reporting of background changes in signal intensity in the prostate that can negatively affect cancer detection; creation of a new category for lesions that do not fit into the PI-RADS assessment categories (i.e., PI-RADS M category); inclusion of quantitative parameters beyond size to evaluate lesion aggressiveness; adjustments to the structured report template, including standardized assessment of the risk of extraprostatic extension; development of parameters for image quality and performance control; and suggestions for expansion of the system to other indications (e.g., active surveillance and recurrence).

Comparison of PI-RADS version 2.1 and PI-RADS version 2 regarding interreader variability and diagnostic accuracy for transition zone prostate cancer

PURPOSE

To compare the diagnostic performance of PI-RADS version 2.1 (PI-RADS v2.1) and PI-RADS v2 for transition zone prostate cancer (TZPC), and analyse its performance for readers with different experience levels.

METHODS

Eighty-five patients with suspected prostate cancer who underwent biopsy after MRI scan between January and December 2017 were retrospectively enrolled. One junior radiologist (reader 1, 1 year of experience in using PI-RADS v2) and one senior radiologist (reader 2, 6 years of experience) independently reviewed and assigned a score for each lesion according to PI-RADS v2.1 and v2. The template-guided transperineal prostate biopsy was used for standard of reference. To compare the diagnostic performance of the two methods, the AUC was calculated. The sensitivity, specificity, and accuracy were calculated at predefined positive values (PI-RADS ≥ 3). The interreader agreement and frequency of prostate cancer for each PI-RADS category were also calculated.

RESULTS

Among the 85 patients, 27 had prostate cancers, and 25 were clinically significant prostate cancer (csPCa). The AUC values for diagnosing clinically significant prostate cancer significantly increased with PI-RADS v2.1 for reader 2 (0.766 vs. 0.902, P = 0.009). The specificity and accuracy for both readers also increased with PI-RADS v2.1 (specificity: reader 1, 41.7% vs. 78.3% and reader 2, 33.3% vs. 81.7%; accuracy: reader 1, 52.9% vs. 76.5% and reader 2, 48.2% vs. 83.5%, all P < 0.05). The interreader agreement was good for both versions. The percentage of prostate cancer decreased in lower PI-RADS categories (PI-RADS 2) and increased in higher PI-RADS categories (PI-RADS 3 ~ 4).

CONCLUSION

Compared with PI-RADS v2, PI-RADS v2.1 may improve radiologists' diagnostic performance for TZPC.

PI-RADS: what is new and how to use it

The prostate imaging reporting and data system (PI-RADS) has revolutionized the use of magnetic resonance imaging (MRI) for the management of prostate cancer (PCa). The most recent version 2.1, PI-RADS v2.1, provides specific refinements in the performance, relaxing some recommendations which were not found to be helpful, while reinforcing and clarifying others. The interpretation of T2-weighted imaging (T2WI) in the transition zone (TZ), and the overall assessment of TZ nodules, now allows for a clearer distinction between those which are clearly benign and those which might warrant tissue sampling. Additional changes also resolve discrepancies in T2WI and diffusion-weighted imaging (DWI) of the peripheral zone (PZ). PI-RADS v2.1 is a simpler, more straightforward, and more reproducible method to better communicate between physicians regarding findings on prostate MRI.

Correlation between Intraprostatic PSMA Uptake and MRI PI-RADS of [68Ga]Ga-PSMA-11 PET/MRI in Patients with Prostate Cancer: Comparison of PI-RADS Version 2.0 and PI-RADS Version 2.1

PURPOSE

We aimed to evaluate the correlation between PSMA uptake and magnetic resonance imaging (MRI) PI-RADS of simultaneous [68Ga]Ga-PSMA-11 PET/MRI regarding PI-RADS version 2.0 and 2.1 respectively and compared the difference between these two versions.

MATERIALS AND METHODS

We retrospectively analyzed a total of forty-six patients with biopsy-proven prostate cancer who underwent simultaneous [68Ga]Ga-PSMA-11 PET/MRI. We classified the lesions regarding PI-RADS version 2.0 and 2.1, peripheral zone (PZ), and transitional zone (TZ), respectively. Based on regions of interest (ROI), standardized uptake values maximum (SUVmax), and corresponding lesion-to-background ratios (LBR) of SUVmax of each category, PI-RADS score 1 to 5, were measured. A comparison between PI-RADS version 2.0 and PI-RADS version 2.1 was performed.

RESULTS

A total of 215 focal prostate lesions were analyzed, including two subgroups, 125 TZ and 90 PZ. Data are reported as median and interquartile range (IQR). Regarding PI-RADS version 2.1, TZ SUVmax of each category were 1.5 (0.5, 1.9), 1.9 (0.8, 2.3), 3.3 (2.1, 4.6), 4.2 (3.1, 5.7), 7.3 (5.2, 9.7). PZ SUVmax of each category were 1.0 (0.8, 1.6), 2.5 (1.5, 3.2), 3.3 (1.9, 4.5), 4.3 (3.0, 5.4), 7.4 (5.0, 9.3). Regarding the inter-reader agreement of the overall PI-RADS assessment category, the kappa value was 0.723 for version 2.0 and 0.853 for version 2.1.

CONCLUSION

Revisions of PI-RADS version 2.1 results in variations in lesions classification. Lesions with the PI-RADS category of 3, 4, and 5 present relatively higher intraprostatic PSMA uptake, while lesions with the PI-RADS category of 1 and 2 present relatively lower and similar uptake. Version 2.1 has higher inter-reader reproducibility than version 2.0.

Shape Analysis of Peripheral Zone Observations on Prostate DWI: Correlation to Histopathology Outcomes After Radical Prostatectomy

OBJECTIVE. The objective of our study was to subjectively and quantitatively assess shape features of peripheral zone (PZ) tumors at DWI compared with pathologic outcomes. MATERIALS AND METHODS. During the study period, 241 consecutive men with PZ dominant prostate tumors underwent 3-T MRI including DWI before undergoing radical prostatectomy. DW images of these patients were retrospectively assessed by two blinded radiologists. The reviewers assigned Prostate Imaging Reporting and Data System (PI-RADS) shape categories (round or oval, crescentic [i.e., conforming to PZ], linear or wedge-shaped) and segmented tumors for quantitative shape analysis. Discrepancies were resolved by consensus. Comparisons were performed with Gleason score (GS) and pathologic stage. RESULTS. Consensus review results were as follows: 63.9% (154/241) of tumors were round or oval; 22.8% (55/241), crescentic; and 13.3% (32/241), linear or wedge-shaped. Agreement for shape assessment was moderate (κ = 0.41). Round or oval tumors were higher grade (GS 6 = 1.3%, GS 7 = 78.0%, GS ≥ 8 = 20.7%) than crescentic tumors (GS 6 = 9.1%, GS 7 = 74.6%, GS ≥ 8 = 16.3%) and linear or wedge-shaped tumors (GS 6 = 6.3%, GS 7 = 78.1%, GS ≥ 8 = 15.6%) (p = 0.011). In addition, round or oval tumors had higher rates of extraprostatic extension (EPE) and seminal vesicle invasion (SVI) (EPE and SVI: 70.1% and 26.0%) than crescentic tumors (67.3% and 9.1%; p = 0.003) and linear or wedge-shaped tumors (40.6% and 9.4%; p = 0.008). Quantitatively, the shape features termed "circularity" and "roundness" were associated with EPE (p < 0.001 and p = 0.003), SVI (p < 0.001 and p = 0.029), and increasing GS (p = 0.009 and p = 0.021), but there was overlap between groups. CONCLUSION. In this study, approximately 10% of resected PZ tumors were linear or wedge-shaped on DWI. PZ tumors that were judged subjectively and evaluated quantitatively to be round or oval were associated with increased prostate cancer aggressiveness.

Validation of the PI-RADS language: predictive values of PI-RADS lexicon descriptors for detection of prostate cancer

OBJECTIVES

To assess the discriminatory power of lexicon terms used in PI-RADS version 2 to describe MRI features of prostate lesions.

METHODS

Four hundred fifty-four patients were included in this retrospective, institutional review board-approved study. Patients received multiparametric (mp) MRI and subsequent prostate biopsy including MRI/transrectal ultrasound fusion biopsy and 10-core systematic biopsy. PI-RADS lexicon terms describing lesion characteristics on mpMRI were assigned to lesions by experienced readers. Positive and negative predictive values (PPV, NPV) of each lexicon term were assessed using biopsy results as a reference standard.

RESULTS

From a total of 501 lesions, clinically significant prostate cancer (csPCa) was present in 175 lesions (34.9%). Terms related to findings of restricted diffusion showed PPVs of up to 52.0%/43.9% and NPV of up to 91.8%/89.7% (peripheral zone or PZ/transition zone or TZ). T2-weighted imaging (T2W)-related terms showed a wide range of predictive values. For PZ lesions, high PPVs were found for "markedly hypointense," "lenticular," "lobulated," and "spiculated" (PPVs between 67.2 and 56.7%). For TZ lesions, high PPVs were found for "water-drop-shaped" and "erased charcoal sign" (78.6% and 61.0%). The terms "encapsulated," "organized chaos," and "linear" showed to be good predictors for benignity with distinctively low PPVs between 5.4 and 6.9%. Most T2WI-related terms showed improved predictive values for TZ lesions when combined with DWI-related findings.

CONCLUSIONS

Lexicon terms with high discriminatory power were identified (e.g., "markedly hypointense," "water-drop-shaped," "organized chaos"). DWI-related terms can be useful for excluding TZ cancer. Combining T2WI- with DWI findings in TZ lesions markedly improved predictive values.

KEY POINTS

• Lexicon terms describing morphological and functional features of prostate lesions on MRI show a wide range of predictive values for prostate cancer. • Some T2-related terms have favorable PPVs, e.g., "water-drop-shaped" and "organized chaos" while others show less distinctive predictive values. DWI-related terms have noticeable negative predictive values in TZ lesions making DWI feature a useful tool for exclusion of TZ cancer. • Combining DWI- and T2-related lexicon terms for assessment of TZ lesions markedly improves PPVs. Most T2-related lexicon terms showed a significant decrease in PPV when combined with negative findings for "DW hyperintensity."

Direct Comparison of PI-RADS Version 2 and 2.1 in Transition Zone Lesions for Detection of Prostate Cancer: Preliminary Experience

BACKGROUND

There appears to be less agreement in the identification of cancers in the transition zone (TZ), which is not as reliable as those in peripheral zone when using the Prostate Imaging Reporting and Data System (PI-RADS) version 2 (v2). In response to such shortcomings, the updated version 2.1 was introduced, which incorporated diffusion-weighted imaging (DWI) into category 2 and clarified lexicons.

PURPOSE

To compare the diagnostic performance for the detection of clinically significant TZ prostate cancers (csPCa) and interreader agreement between PI-RADS v2.1 and v2.

STUDY TYPE

Retrospective study.

POPULATION

In all, 142 patients, 201 TZ lesions.

FIELD STRENGTH/SEQUENCE

3.0T; T₂ -weighted image and DWI.

ASSESSMENT

Lesions were scored by three independent readers using PI-RADS v2 and v2.1.

STATISTICAL TESTS

The sensitivity and specificity at category ≥3 were compared between v2 and v2.1 using the generalized estimating equation model. Detection rates for csPCa of upgraded and downgraded lesions in the use of PI-RADS v2.1 from v2 were assessed. Interreader agreement was assessed using κ statistics.

RESULTS

PI-RADS v2.1 showed a higher sensitivity and specificity (94.5% and 60.9%) than v2 (91.8% and 56.3%) for category ≥3 lesions in the detection of csPCa, although not significantly. Of eight upgraded lesions from category 2 to 3 (2 + 1) with an incorporated DWI, 50% (4/8) were csPCa. This was significantly higher than category 2 lesions (4.4%; P = 0.003). No csPCa was detected among the 22.8% (46/201) downgraded lesions. There was a moderate interreader agreement for scores ≥3 (κ = 0.565) in v2.1, which was slightly higher than that for v2 (κ = 0.534), although not significantly.

DATA CONCLUSION

PI-RADS v2.1 provides moderate and comparable interreader agreement at category ≥3 than v2 in the TZ lesions. Upgraded lesions from category 2 to 3 demonstrated a higher detection rate of csPCa than category 2 lesions in v2.1.

LEVEL OF EVIDENCE

4 TECHNICAL EFFICACY STAGE: 2 J. Magn. Reson. Imaging 2020;52:577-586.

Comparison of PI-RADS version 2 and PI-RADS version 2.1 for the detection of transition zone prostate cancer

PURPOSE

To compare the diagnostic performance of PI-RADS v2 and v2.1 for detecting transition zone prostate cancer (TZPC) on multiparametric prostate MRI (mpMRI).

METHOD

Fifty-eight patients with elevated PSA levels underwent mpMRI at 3 T including T2-weighted imaging (T2WI) and diffusion-weighted imaging (DWI), and subsequent MRI-transrectal ultrasonography fusion-guided prostate-targeted biopsy (MRGB). The standard of reference was MRGB-derived histopathology. Two readers independently assessed each TZ lesion, assigning a score of 1-5 for T2WI, a score of 1-5 for DWI, and the overall PI-RADS assessment category according to PI-RADS v2 and v2.1. The diagnostic performance of the two methods was compared in terms of inter-reader agreement, diagnostic sensitivity, diagnostic specificity, and area under the ROC curve (AUC).

RESULTS

Of the 58 patients, 26 were diagnosed with PC (GS = 3 + 3, n = 9; GS = 3 + 4, n = 9; GS = 3 + 5, n = 1; GS = 4 + 3, n = 4; GS = 4 + 4, n = 3) and 32 with benign lesions. Regarding inter-reader agreement of overall PI-RADS assessment category, the kappa value was 0.580 for v2 and 0.645 for v2.1. For both readers, there was no difference in diagnostic sensitivity between the versions (p ≥ 0.500). For reader 1, the diagnostic specificity was higher for v2.1 (p = 0.002), and was similar for reader 2 (p = 1.000). For both readers, AUC tended to be higher for v2.1 than for v2, but the difference was not significant (0.786 vs. 0.847 for reader 1, p = 0.052; and 0.808 vs. 0.858 for reader 2, p = 0.197).

CONCLUSIONS

These results suggest that compared with PI-RADS v2, PI-RADS v2.1 could be preferable for evaluating TZ lesions.

PI-RADS version 2.1: one small step for prostate MRI

Multiparametric (mp) prostate magnetic resonance imaging (MRI) is playing an increasingly prominent role in the diagnostic work-up of patients with suspected prostate cancer. Performing mpMRI before biopsy offers several advantages including biopsy avoidance under certain clinical circumstances and targeting biopsy of suspicious lesions to enable the correct diagnosis. The success of the technique is heavily dependent on high-quality image acquisition, interpretation, and report communication, all areas addressed by previous versions of the Prostate Imaging-Reporting and Data System (PI-RADS) recommendations. Numerous studies have validated the approach, but the widespread adoption of PI-RADS version 2 has also highlighted inconsistencies and limitations, particularly relating to interobserver variability for evaluation of the transition zone. These limitations are addressed in the recently released version 2.1. In this article, we highlight the key changes proposed in PI-RADS v2.1 and explore the background reasoning and evidence for the recommendations.

Interreader Variability of Prostate Imaging Reporting and Data System Version 2 in Detecting and Assessing Prostate Cancer Lesions at Prostate MRI

OBJECTIVE. The purpose of this study was to evaluate agreement among radiologists in detecting and assessing prostate cancer at multiparametric MRI using Prostate Imaging Reporting and Data System version 2 (PI-RADSv2). MATERIALS AND METHODS. Treatment-naïve patients underwent 3-T multipara-metric MRI between April 2012 and June 2015. Among the 163 patients evaluated, 110 underwent prostatectomy after MRI and 53 had normal MRI findings and transrectal ultrasound-guided biopsy results. Nine radiologists participated (three each with high, intermediate, and low levels of experience). Readers interpreted images of 58 patients on average (range, 56-60) using PI-RADSv2. Prostatectomy specimens registered to MRI were ground truth. Interob-server agreement was evaluated with the index of specific agreement for lesion detection and kappa and proportion of agreement for PI-RADS category assignment. RESULTS. The radiologists detected 336 lesions. Sensitivity for index lesions was 80.9% (95% CI, 75.1-85.9%), comparable across reader experience (p = 0.392). Patient-level specificity was experience dependent; highly experienced readers had 84.0% specificity versus 55.2% for all others (p < 0.001). Interobserver agreement was excellent for detecting index lesions (index of specific agreement, 0.871; 95% CI, 0.798-0.923). Agreement on PI-RADSv2 category assignment of index lesions was moderate (κ = 0.419; 95% CI, 0.238-0.595). For individual category assignments, proportion of agreement was slight for PI-RADS category 3 (0.208; 95% CI, 0.086-0.284) but substantial for PI-RADS category 4 (0.674; 95% CI, 0.540-0.776). However, proportion of agreement for T2-weighted PI-RADS 4 in the transition zone was 0.250 (95% CI, 0.108-0.372). Proportion of agreement for category assignment of index lesions on dynamic contrast-enhanced MR images was 0.822 (95% CI, 0.728-0.903), on T2-weighted MR images was 0.515 (95% CI, 0.430-0623), and on DW images was 0.586 (95% CI, 0.495-0.682). Proportion of agreement for dominant lesion was excellent (0.828; 95% CI, 0.742-0.913). CONCLUSION. Radiologists across experience levels had excellent agreement for detecting index lesions and moderate agreement for category assignment of lesions using PI-RADS. Future iterations of PI-RADS should clarify PI-RADS 3 and PI-RADS 4 in the transition zone.

Prostate Imaging-Reporting and Data System Steering Committee: PI-RADS v2 Status Update and Future Directions

CONTEXT

The Prostate Imaging-Reporting and Data System (PI-RADS) v2 analysis system for multiparametric magnetic resonance imaging (mpMRI) detection of prostate cancer (PCa) is based on PI-RADS v1, accumulated scientific evidence, and expert consensus opinion.

OBJECTIVE

To summarize the accuracy, strengths and weaknesses of PI-RADS v2, discuss pathway implications of its use and outline opportunities for improvements and future developments.

EVIDENCE ACQUISITION

For this consensus expert opinion from the PI-RADS steering committee, clinical studies, systematic reviews, and professional guidelines for mpMRI PCa detection were evaluated. We focused on the performance characteristics of PI-RADS v2, comparing data to systems based on clinicoradiologic Likert scales and non-PI-RADS v2 imaging only. Evidence selections were based on high-quality, prospective, histologically verified data, with minimal patient selection and verifications biases.

EVIDENCE SYNTHESIS

It has been shown that the test performance of PI-RADS v2 in research and clinical practice retains higher accuracy over systematic transrectal ultrasound (TRUS) biopsies for PCa diagnosis. PI-RADS v2 fails to detect all cancers but does detect the majority of tumors capable of causing patient harm, which should not be missed. Test performance depends on the definition and prevalence of clinically significant disease. Good performance can be attained in practice when the quality of the diagnostic process can be assured, together with joint working of robustly trained radiologists and urologists, conducting biopsy procedures within multidisciplinary teams.

CONCLUSIONS

It has been shown that the test performance of PI-RADS v2 in research and clinical practice is improved, retaining higher accuracy over systematic TRUS biopsies for PCa diagnosis.

PATIENT SUMMARY

Multiparametric magnetic resonance imaging (MRI) and MRI-directed biopsies using the Prostate Imaging-Reporting and Data System improves the detection of prostate cancers likely to cause harm, and at the same time decreases the detection of disease that does not lead to harms if left untreated. The keys to success are high-quality imaging, reporting, and biopsies by radiologists and urologists working together in multidisciplinary teams.

Diagnostic performance of diffusion-weighted imaging for prostate cancer: Peripheral zone versus transition zone

OBJECTIVES

Diffusion-weighted imaging (DWI) has been shown to be an important component of multiparametric magnetic resonance imaging (mpMRI). We compared performance of DWI for detection of prostate cancer (PCa) in peripheral zone (PZ) and transition zone (TZ) of prostate.

MATERIALS AND METHODS

We reviewed data of 460 subjects who underwent preoperative 3.0-Tesla mpMRI and subsequently radical prostatectomy. Level of suspicion for PCa was graded using 5-grade Likert-scale from DWI. Topographic analyses were performed for location of tumor foci at each surgical specimen. Among those with DWI grade ≥ III, we analyzed concordance rate on the location of radiologic and pathologic index lesions between DWI and surgical specimens.

RESULTS

Among 460 patients, 351 (76.3%) patients showed suspicious DWI lesions (57.5% in PZ, 42.5% in TZ). Multivariates regression analyses revealed significant associations between high DWI grade and adverse pathologic outcomes including pathologic stage, Gleason score, tumor volume and extracapsular extension (all p < 0.05). Overall concordance rates between DWI and surgical specimen were 75.8%, significantly higher in PZ than TZ (82.2% vs. 67.1% p = 0.002). Such concordance rate showed a positive linear association with increase in DWI grading (p < 0.001). Among 109 patients with DWI grade I-II, 28 (25.7%) harbored high grade disease (pathologic Gleason score ≥ 4 + 3).

CONCLUSIONS

DWI detects tumors in PZ of prostate more accurately than those in TZ. Such accuracy of DWI was shown to be more evident with higher DWI grade. Meanwhile, a negative DWI did not guarantee absence of high grade PCa.

Accuracy of multiparametric MR imaging with PI-RADS V2 assessment in detecting infiltration of the neurovascular bundles prior to prostatectomy

OBJECTIVES

To evaluate the accuracy of assessment of neurovascular bundle (NVB) infiltration using multiparametric magnetic resonance imaging (mpMRI) and PI-RADS V2 prior to prostatectomy.

METHODS

The ethics committee approved this retrospective study with waiver of informed consent. N=198 consecutive patients with biopsy proved cancer underwent standardized mpMRI at 3T prior to surgery. NVB infiltration was assessed for each side (a total of 396). Maximum PI-RADS V2 scores were determined for the posterolateral areas adjacent to the NVBs. Imaging results were correlated with postoperative pathology and standard descriptive statistics were calculated.

RESULTS

Overall T-staging sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and accuracy of mpMRI were 64.4%, 89.2%, 82.4%, 76.2% and 78.3%, respectively. In 396 cases NVB infiltration was predicted with 75.3%, 94.0%, 80.2%, 92.1 % and 89.4 % sensitivity, specificity, PPV, NPV and accuracy, respectively. Analyses of 396 NVB and their adjacent PI-RADS V2 scores with pathology revealed significantly more NVB-infiltrations in suspect scores of 5 and 4 vs. uncertain scores of 3-1 (81/264 vs. 16/132, p=0.0001). Considering scores higher than 3 as a criterion of infiltration demonstrated moderate sensitivity and poor specificity (83.5% and 38.8%, respectively). Interobserver agreement of a second reading of a random sample was good (κ=0.64) for NVB infiltrations and moderate (κ=0.59) for PI-RADS V2.

CONCLUSIONS

Assessment of infiltration of the neurovascular bundles using mpMRI has valuable diagnostic performance, yet PI-RADS V2 Scores demonstrate limited eligibility. Combined findings offer crucial information for the planning of prostatectomy.

Support Vector Machines (SVM) classification of prostate cancer Gleason score in central gland using multiparametric magnetic resonance images: A cross-validated study

PURPOSE

To assess the performance of Support Vector Machines (SVM) classification to stratify the Gleason Score (GS) of prostate cancer (PCa) in the central gland (CG) based on image features across multiparametric magnetic resonance imaging (mpMRI).

MATERIALS AND METHODS

This retrospective study was approved by the institutional review board, and informed consent was waived. One hundred fifty-two CG cancerous ROIs were identified through radiological-pathological correlation. Eleven parameters were derived from the mpMRI and histogram analysis, including mean, median, the 10th percentile, skewness and kurtosis, was performed for each parameter. In total, fifty-five variables were calculated and processed in the SVM classification. The classification model was developed with 10-fold cross-validation and was further validated mutually across two separated datasets.

RESULTS

With six variables selected by a feature-selection and variation test, the prediction model yielded an area under the receiver operating characteristics curve (AUC) of 0.99 (95% CI: 0.98, 1.00) when trained in dataset A2 and 0.91 (95% CI: 0.85, 0.95) for the validation in dataset B2. When the data sets were reversed, an AUC of 0.99 (95% CI: 0.99, 1.00) was obtained when the model was trained in dataset B2 and 0.90 (95% CI: 0.85, 0.95) for the validation in dataset A2.

CONCLUSION

The SVM classification based on mpMRI derived image features obtains consistently accurate classification of the GS of PCa in the CG.