Search

Simple Summary: The proper management of prostate cancer requires accurate assessment of patients diagnosed with prostate cancer, a common and often lethal cancer. Current conventional prostate cancer diagnosis relies on: (1) needle biopsy, which can result in side effects such as hemorrhage or infection, as well as misplacement of needles resulting in evaluation inaccuracy, (2) prostate serum antigen detection that can lead to overdiagnosis, or (3) MRI interpretations using the Prostate Imaging-Reporting and Data System (PI-RADS), which is a subjective method that depends on the experience of the radiologist's interpretation of the images. Quantitative analysis of MRI can alleviate the limitations of conventional approaches. Artificial intelligence applied to MRI requires big training datasets, and its application is restricted to patients scanned under certain restricted conditions. More recently, spectral/statistical techniques, adapted from remote sensing, have been applied and tested in spatially registered multi-parametric MRI. The new spectral/statistical techniques require limited training, are flexible regarding patient scanning conditions, and perform well relative to other techniques. This study extends spectral/statistical techniques to make them more independent of the individual radiologist by autonomously finding the appropriate tumor signature within a prostate and, therefore, providing an objective, non-invasive, accurate means for determining prostate tumor aggressiveness. Background: Accurate, reliable, non-invasive assessment of patients diagnosed with prostate cancer is essential for proper disease management. Quantitative assessment of multi-parametric MRI, such as through artificial intelligence or spectral/statistical approaches, can provide a non-invasive objective determination of the prostate tumor aggressiveness without side effects or potential poor sampling from needle biopsy or overdiagnosis from prostate serum antigen measurements. To simplify and expedite prostate tumor evaluation, this study examined the efficacy of autonomously extracting tumor spectral signatures for spectral/statistical algorithms for spatially registered bi-parametric MRI. Methods: Spatially registered hypercubes were digitally constructed by resizing, translating, and cropping from the image sequences (Apparent Diffusion Coefficient (ADC), High B-value, T2) from 42 consecutive patients in the bi-parametric MRI PI-CAI dataset. Prostate cancer blobs exceeded a threshold applied to the registered set from normalizing the registered set into an image that maximizes High B-value, but minimizes the ADC and T2 images, appearing "green" in the color composite. Clinically significant blobs were selected based on size, average normalized green value, sliding window statistics within a blob, and position within the hypercube. The center of mass and maximized sliding window statistics within the blobs identified voxels associated with tumor signatures. We used correlation coefficients (R) and p-values, to evaluate the linear regression fits of the z-score and SCR (with processed covariance matrix) to tumor aggressiveness, as well as Area Under the Curves (AUC) for Receiver Operator Curves (ROC) from logistic probability fits to clinically significant prostate cancer. Results: The highest R (R > 0.45), AUC (>0.90), and lowest p-values (<0.01) were achieved using z-score and modified registration applied to the covariance matrix and tumor signatures selected from the "greenest" parts from the selected blob. Conclusions: The first autonomous tumor signature applied to spatially registered bi-parametric MRI shows promise for determining prostate tumor aggressiveness. [ABSTRACT FROM AUTHOR]

Introduction: Accurate diagnosis and treatment of kidney tumors greatly benefit from automated solutions for detection and classification on MRI. In this study, we explore the application of a deep learning algorithm, YOLOv7, for detecting kidney tumors on contrast-enhanced MRI. Material and methods: We assessed the performance of YOLOv7 tumor detection on excretory phase MRIs in a large institutional cohort of patients with RCC. Tumors were segmented on MRI using ITK-SNAP and converted to bounding boxes. The cohort was randomly divided into ten benchmarks for training and testing the YOLOv7 algorithm. The model was evaluated using both 2-dimensional and a novel in-house developed 2.5-dimensional approach. Performance measures included F1, Positive Predictive Value (PPV), Sensitivity, F1 curve, PPV-Sensitivity curve, Intersection over Union (IoU), and mean average PPV (mAP). Results: A total of 326 patients with 1034 tumors with 7 different pathologies were analyzed across ten benchmarks. The average 2D evaluation results were as follows: Positive Predictive Value (PPV) of 0.69 ± 0.05, sensitivity of 0.39 ± 0.02, and F1 score of 0.43 ± 0.03. For the 2.5D evaluation, the average results included a PPV of 0.72 ± 0.06, sensitivity of 0.61 ± 0.06, and F1 score of 0.66 ± 0.04. The best model performance demonstrated a 2.5D PPV of 0.75, sensitivity of 0.69, and F1 score of 0.72. Conclusion: Using computer vision for tumor identification is a cutting-edge and rapidly expanding subject. In this work, we showed that YOLOv7 can be utilized in the detection of kidney cancers. [ABSTRACT FROM AUTHOR]

Introduction: Classification of clear cell renal cell carcinoma (ccRCC) growth rates in patients with Von Hippel–Lindau (VHL) syndrome has several ramifications for tumor monitoring and surgical planning. Using two separate machine-learning algorithms, we sought to produce models to predict ccRCC growth rate classes based on qualitative MRI-derived characteristics. Material and Methods: We used a prospectively maintained database of patients with VHL who underwent surgical resection for ccRCC between January 2015 and June 2022. We employed a threshold growth rate of 0.5 cm per year to categorize ccRCC tumors into two distinct groups—'slow-growing' and 'fast-growing'. Utilizing a questionnaire of qualitative imaging features, two radiologists assessed each lesion on different MRI sequences. Two machine-learning models, a stacked ensemble technique and a decision tree algorithm, were used to predict the tumor growth rate classes. Positive predictive value (PPV), sensitivity, and F1-score were used to evaluate the performance of the models. Results: This study comprises 55 patients with VHL with 128 ccRCC tumors. Patients' median age was 48 years, and 28 patients were males. Each patient had an average of two tumors, with a median size of 2.1 cm and a median growth rate of 0.35 cm/year. The overall performance of the stacked and DT model had 0.77 ± 0.05 and 0.71 ± 0.06 accuracies, respectively. The best stacked model achieved a PPV of 0.92, a sensitivity of 0.91, and an F1-score of 0.90. Conclusion: This study provides valuable insight into the potential of machine-learning analysis for the determination of renal tumor growth rate in patients with VHL. This finding could be utilized as an assistive tool for the individualized screening and follow-up of this population. [ABSTRACT FROM AUTHOR]

Men of African ancestry have a disproportionately higher incidence of prostate cancer and likelihood of mortality when compared to men of European ancestry. The underlying reasons for this are unknown and may be due to an interplay of genomic and environmental factors. In this patient population, higher incidence can potentially be counteracted by earlier screening and more intense surveillance. Recent American Urological Association guidelines now support screening men of African ancestry beginning at age 40-45 years with a grade B recommendation. Multiparametric MRI (mpMRI) has proven to be very useful in detecting clinically significant prostate cancer, but access to this resource is often limited among communities of color. In this clinical trial, we will assess the influence of mpMRI and MRI/US fusion-guided biopsy on detection rates of clinically significant prostate cancer in men of African ancestry and compare these results with self-referred patients who are predominantly of European origin. This clinical trial (NCT05872503) is a prospective, single-arm observational study. Up to 345 biopsy-naïve patients age >35 years with PSA ≥3ng/mL, positive rectal exam, or strong family history of prostate cancer who self-identify as having African ancestry will be recruited. Study patients will be referred by local hospitals that provide care mainly to communities of color, and patients will undergo mpMRI at the National Cancer Institute. Patients with identified PI-RADS ≥3 lesions will undergo MRI/US fusion-guided biopsy within six months of the initial mpMRI and will receive standard-of-care treatment and follow-up for five years with annual phone or virtual visits. Patients without PI-RADS ≥3 lesions will not receive biopsy and will be followed for five years via serum PSA and mpMRI. Cancer detection rates at the patient-level between men of African ancestry and previously accrued patients of European ancestry will be compared using Fisher's exact test of independence. Results: Conclusions: [ABSTRACT FROM AUTHOR]

The impact of biopsy approach on the detection rate of prostate cancer (PCa) is an active area of investigation. While recent evidence suggests higher cancer detection rates for targeted biopsies using a transperineal approach, the impact of lesion location remains unclear. This study aims to investigate the relative performance of transperineal (TP) MRI-targeted prostate biopsy (TBx) for the detection of clinically significant prostate cancer (csPCa) in the apex, compared to transrectal (TR) TBx. A prospectively maintained database of patients who underwent MRI-guided prostate biopsy at our institution between 2016-2023 was queried for patients with MRI lesions in the apex.;Biopsies were performed using the UroNav MRI/ultrasound fusion biopsy system using an endfire ultrasound probe. Overall cancer detection rate, as well as rate of Gleason grade group (GG) ≥2 and GG≥3, were compared on a per-lesion basis between patients undergoing TP-TBx and patients undergoing TR-TBx. Clinical factors including PSA levels, prostate volume, and the location of the lesion in the apex were assessed using two-sample t-tests and chi-squared tests, and PIRADS scores were compared between groups using the Kruskal-Wallis test. To account for patients who underwent both TP-TBx and TR-TBx, chi-squared tests were conducted to compare the diagnoses for the same patient cohort. Between 2016-2023, 1034 apical prostate lesion biopsies were identified, of which 159 were TP-TBx and 875 were TR-TBx (Table 1). A univariate analysis of pre-biopsy PSA, prostate volume and PIRADS demonstrated no statistically significant difference between the TP-TBx and TR-TBx groups. The rate of ≥GG2 cancers diagnosed was statistically greater for TP-TBx vs TR-TBx (44.26% vs 35.66%, p =0.03). The overall cancer detection rate (CDR) and CDR of cancers ≥ GG 3 were not statistically different between biopsy types. Amongst the sub cohort with available zonal delineation (n = 764), there was a higher CDR of GG>2 cancers in the peripheral zone, but not transition zone (Table 2). Among a subset of patients who underwent TR biopsy followed by TP (n= 34), 35.29% were upgraded on TP biopsy, while 0% of patients who underwent TR after TP biopsy (n= 5) were upgraded on TR biopsy. In this analysis, targeted transperineal biopsy was associated with higher GG≥2 detection rate compared to transrectal biopsy for apical lesions. Prospective studies are warranted to validate transperineal biopsy as a potentially superior approach for prostate cancer detection in apical lesions in the MRI era. [ABSTRACT FROM AUTHOR]

Objective: To assess impact of image quality on prostate cancer extraprostatic extension (EPE) detection on MRI using a deep learning-based AI algorithm.This retrospective, single institution study included patients who were imaged with mpMRI and subsequently underwent radical prostatectomy from June 2007 to August 2022. One genitourinary radiologist prospectively evaluated each patient using the NCI EPE grading system. Each T2WI was classified as low- or high-quality by a previously developed AI algorithm. Fisher’s exact tests were performed to compare EPE detection metrics between low- and high-quality images. Univariable and multivariable analyses were conducted to assess the predictive value of image quality for pathological EPE.A total of 773 consecutive patients (median age 61 [IQR 56–67] years) were evaluated. At radical prostatectomy, 23% (180/773) of patients had EPE at pathology, and 41% (131/318) of positive EPE calls on mpMRI were confirmed to have EPE. The AI algorithm classified 36% (280/773) of T2WIs as low-quality and 64% (493/773) as high-quality. For EPE grade ≥ 1, high-quality T2WI significantly improved specificity for EPE detection (72% [95% CI 67–76%] vs. 63% [95% CI 56–69%], P = 0.03), but did not significantly affect sensitivity (72% [95% CI 62–80%] vs. 75% [95% CI 63–85%]), positive predictive value (44% [95% CI 39–49%] vs. 38% [95% CI 32–43%]), or negative predictive value (89% [95% CI 86–92%] vs. 89% [95% CI 85–93%]). Sensitivity, specificity, PPV, and NPV for EPE grades ≥ 2 and ≥ 3 did not show significant differences attributable to imaging quality. For NCI EPE grade 1, high-quality images (OR 3.05, 95% CI 1.54–5.86; P < 0.001) demonstrated a stronger association with pathologic EPE than low-quality images (OR 1.76, 95% CI 0.63–4.24; P = 0.24).Our study successfully employed a deep learning-based AI algorithm to classify image quality of prostate MRI and demonstrated that better quality T2WI was associated with more accurate prediction of EPE at final pathology.Materials and methods: To assess impact of image quality on prostate cancer extraprostatic extension (EPE) detection on MRI using a deep learning-based AI algorithm.This retrospective, single institution study included patients who were imaged with mpMRI and subsequently underwent radical prostatectomy from June 2007 to August 2022. One genitourinary radiologist prospectively evaluated each patient using the NCI EPE grading system. Each T2WI was classified as low- or high-quality by a previously developed AI algorithm. Fisher’s exact tests were performed to compare EPE detection metrics between low- and high-quality images. Univariable and multivariable analyses were conducted to assess the predictive value of image quality for pathological EPE.A total of 773 consecutive patients (median age 61 [IQR 56–67] years) were evaluated. At radical prostatectomy, 23% (180/773) of patients had EPE at pathology, and 41% (131/318) of positive EPE calls on mpMRI were confirmed to have EPE. The AI algorithm classified 36% (280/773) of T2WIs as low-quality and 64% (493/773) as high-quality. For EPE grade ≥ 1, high-quality T2WI significantly improved specificity for EPE detection (72% [95% CI 67–76%] vs. 63% [95% CI 56–69%], P = 0.03), but did not significantly affect sensitivity (72% [95% CI 62–80%] vs. 75% [95% CI 63–85%]), positive predictive value (44% [95% CI 39–49%] vs. 38% [95% CI 32–43%]), or negative predictive value (89% [95% CI 86–92%] vs. 89% [95% CI 85–93%]). Sensitivity, specificity, PPV, and NPV for EPE grades ≥ 2 and ≥ 3 did not show significant differences attributable to imaging quality. For NCI EPE grade 1, high-quality images (OR 3.05, 95% CI 1.54–5.86; P < 0.001) demonstrated a stronger association with pathologic EPE than low-quality images (OR 1.76, 95% CI 0.63–4.24; P = 0.24).Our study successfully employed a deep learning-based AI algorithm to classify image quality of prostate MRI and demonstrated that better quality T2WI was associated with more accurate prediction of EPE at final pathology.Results: To assess impact of image quality on prostate cancer extraprostatic extension (EPE) detection on MRI using a deep learning-based AI algorithm.This retrospective, single institution study included patients who were imaged with mpMRI and subsequently underwent radical prostatectomy from June 2007 to August 2022. One genitourinary radiologist prospectively evaluated each patient using the NCI EPE grading system. Each T2WI was classified as low- or high-quality by a previously developed AI algorithm. Fisher’s exact tests were performed to compare EPE detection metrics between low- and high-quality images. Univariable and multivariable analyses were conducted to assess the predictive value of image quality for pathological EPE.A total of 773 consecutive patients (median age 61 [IQR 56–67] years) were evaluated. At radical prostatectomy, 23% (180/773) of patients had EPE at pathology, and 41% (131/318) of positive EPE calls on mpMRI were confirmed to have EPE. The AI algorithm classified 36% (280/773) of T2WIs as low-quality and 64% (493/773) as high-quality. For EPE grade ≥ 1, high-quality T2WI significantly improved specificity for EPE detection (72% [95% CI 67–76%] vs. 63% [95% CI 56–69%], P = 0.03), but did not significantly affect sensitivity (72% [95% CI 62–80%] vs. 75% [95% CI 63–85%]), positive predictive value (44% [95% CI 39–49%] vs. 38% [95% CI 32–43%]), or negative predictive value (89% [95% CI 86–92%] vs. 89% [95% CI 85–93%]). Sensitivity, specificity, PPV, and NPV for EPE grades ≥ 2 and ≥ 3 did not show significant differences attributable to imaging quality. For NCI EPE grade 1, high-quality images (OR 3.05, 95% CI 1.54–5.86; P < 0.001) demonstrated a stronger association with pathologic EPE than low-quality images (OR 1.76, 95% CI 0.63–4.24; P = 0.24).Our study successfully employed a deep learning-based AI algorithm to classify image quality of prostate MRI and demonstrated that better quality T2WI was associated with more accurate prediction of EPE at final pathology.Conclusion: To assess impact of image quality on prostate cancer extraprostatic extension (EPE) detection on MRI using a deep learning-based AI algorithm.This retrospective, single institution study included patients who were imaged with mpMRI and subsequently underwent radical prostatectomy from June 2007 to August 2022. One genitourinary radiologist prospectively evaluated each patient using the NCI EPE grading system. Each T2WI was classified as low- or high-quality by a previously developed AI algorithm. Fisher’s exact tests were performed to compare EPE detection metrics between low- and high-quality images. Univariable and multivariable analyses were conducted to assess the predictive value of image quality for pathological EPE.A total of 773 consecutive patients (median age 61 [IQR 56–67] years) were evaluated. At radical prostatectomy, 23% (180/773) of patients had EPE at pathology, and 41% (131/318) of positive EPE calls on mpMRI were confirmed to have EPE. The AI algorithm classified 36% (280/773) of T2WIs as low-quality and 64% (493/773) as high-quality. For EPE grade ≥ 1, high-quality T2WI significantly improved specificity for EPE detection (72% [95% CI 67–76%] vs. 63% [95% CI 56–69%], P = 0.03), but did not significantly affect sensitivity (72% [95% CI 62–80%] vs. 75% [95% CI 63–85%]), positive predictive value (44% [95% CI 39–49%] vs. 38% [95% CI 32–43%]), or negative predictive value (89% [95% CI 86–92%] vs. 89% [95% CI 85–93%]). Sensitivity, specificity, PPV, and NPV for EPE grades ≥ 2 and ≥ 3 did not show significant differences attributable to imaging quality. For NCI EPE grade 1, high-quality images (OR 3.05, 95% CI 1.54–5.86; P < 0.001) demonstrated a stronger association with pathologic EPE than low-quality images (OR 1.76, 95% CI 0.63–4.24; P = 0.24).Our study successfully employed a deep learning-based AI algorithm to classify image quality of prostate MRI and demonstrated that better quality T2WI was associated with more accurate prediction of EPE at final pathology. [ABSTRACT FROM AUTHOR]

Multiparametric magnetic resonance imaging (mpMRI) has led to improved diagnosis and risk stratification of prostate cancer. The Prostate Imaging Reporting and Data System (PI-RADS) is a widely-adopted standardized interpretation scheme for prostate mpMRI. Lesions are scored 1-5 based on suspicion for clinically significant prostate (csPCa), defined as Grade Group ≥ 2, with increasing positive predictive value indicated by higher score. Multiple studies have demonstrated the positive predictive value of a PI-RADS 5 lesion for csPCa to cluster around 70-80%. Since nearly 80% of those with PI-RADS 5 lesions harbor clinically significant disease on biopsy, many receive upfront therapy. There is limited understanding of what happens to remaining patients with PI-RADS 5 lesions in whom biopsies reveal less aggressive disease. The purpose of this study is to evaluate the clinical course of those with PI-RADS 5 lesions with Grade Group 1 disease who were placed on Active Surveillance (AS). This analysis was limited to AS patients with GG1 prostate cancer on targeted biopsy of a solitary PI-RADS 5 index lesion performed from 2007 to 2022. Each scan was read by a high-volume genitourinary radiologist (BT or PC). These patients underwent systematic and targeted biopsy by physicians experienced with MRI-fusion biopsies (PP, BW, SG). Repeat biopsies were performed based on clinicopathologic concern for progression (rising PSA, MRI changes) or local AS protocols. For MRIs preceding PI-RADS scoring, an internal Likert suspicion scoring system that has been validated in the literature was used and converted into PI-RADS scores. Beginning in 2015, internal mpMRI readings were reported using PI-RADS v2.0 and in 2019 transitioned to PI-RADS v2.1. Biopsy specimens were reviewed by dedicated genitourinary pathologists. Descriptive statistics were performed. Univariate analyses, including independent samples t-test, chi-squared analysis, and Fisher's exact test, were performed to evaluate for factors associated with disease upgrading. 21 patients met inclusion criteria. The mean PSA was 7.9 ng/ml (range 2.4-14.5) (Table 1). At median follow up of 4.3 years, 15/21 (71%) patients progressed to GG≥2 disease, 5 of whom (24% of initial cohort) progressed to GG≥3 (Figure 1). Median time to progression to GG≥2 and GG≥3 was 3.3 years and 4.4 years, respectively. Of the 6 who did not progress, 1 underwent prostatectomy (upgraded to GG2 on final pathology) and 5 were maintained on AS. Of the 16 who progressed to GG≥2, 6/21 (29%) were continued on AS. 5 of these patients remained on AS (median follow-up 3.3 years after progression) and one underwent prostatectomy. Overall, 11 (52%) received definitive therapy. On univariate analysis, only mean maximum cancer core length was associated with progression to radical therapy (4.8 vs 8.0 mm, p=0.048). Patients with PI-RADS 5 lesions and GG1 PCa on AS demonstrate a rate of upgrading to GG≥2 of >70% and nearly a quarter were upgraded to GG≥3. Nearly 50% of patients progressed to radical treatment at 4 years. This represents a high rate of upgrading in a short period of time relative to larger AS series, which place pathologic upgrading at 25-50%. Further studies, including longer-term follow-up, will be necessary. Genomic classification of these initially indolent lesions may provide clarity regarding the best management. The question of how our findings should affect practice is important. With half of patients suitable for AS at follow-up, it is reasonable to continue AS as the preferred management. However, there is a higher risk of requiring intervention and these patients should be counseled and monitored accordingly. While focal therapy is not guideline-endorsed for GG1 disease, this may be a population that could benefit. [ABSTRACT FROM AUTHOR]

The advent of multiparametric magnetic resonance imaging (mpMRI) has helped to localize clinically significant prostate cancer (csPCA). Focal Therapy (FT) has emerged in the mpMRI era, which can treat MRI-visible lesions while avoiding side effects of whole-gland treatment. However, FT remains an investigational treatment option for PCa patients with intermediate-risk disease. Little is known about how many patients on active surveillance (AS) with Gleason grade group (GG) 2 disease retain eligibility for FT over time, or risk factors for loss of eligibility. The objective of this study is to evaluate initial and long-term FT eligibility (FTE) of patients with GG2 disease. A prospectively maintained cohort was retrospectively queried for patients initiated on AS between 2007;and 2020 with GG2 PCa. Patients with biopsy-concordant GG2, unilateral, MRI-visible PIRADS 2-5 lesions amenable to hemiablation were considered FT candidates. Those with PSA>20 ng/mL, >4 MRI lesions, bilateral (BL) MRI-visible/biopsy-concordant lesions, or contralateral GG≥2 PCa or MRI-invisible lesions were excluded. Patients were considered to progress if they developed BL GG≥2 disease or unilateral GG≥3 PCa. Patients without GG progression maintained FTE and those with GG progression maintained FTE if they had unilateral biopsy-concordant/MRI-visible GG≤3, no contralateral GG≥2 disease, and PSA<20 ng/mL. Univariate and multivariate analyses were conducted to compare patients who maintained and lost FTE. 252 PCa patients were identified as eligible for FT with an average follow-up of 3.4 years, 83 of whom (33%) had GG2 disease and were FT candidates. 29/83 (35%) lost FT eligibility, 19/83 (23%) with progression to high-risk disease, 6/83 (7.2%) due to development of BL GG≥2 disease, and 4/83 (4.8%) with development of MRI-invisible csPCA. Baseline characteristics are shown in Table 1. Univariate analysis demonstrated that those who lost FTE had smaller prostates than those who maintained it (44.9 mL vs. 56.2mL, respectively, p=0.018). PSA density was similar between the groups. No significant differences were noted in initial imaging and pathologic parameters. No factors significantly associated with loss of FTE were identified on logistic regression (Table 2). 33% of the patients on AS had GG2 disease and were initially eligible for FT. 65% of patients with GG2 disease were still eligible for FT after a median follow up of 3.4 years; over one third of these patients lose FTE. This data regarding the loss of a FT therapeutic window could prove useful in counseling GG2 patients considering AS, FT, or radical treatment. [ABSTRACT FROM AUTHOR]

The introduction of multiparametric magnetic resonance imaging (mpMRI) has significantly improved the early detection of clinically significant prostate cancer. The use of mpMRI has enabled more precise disease characterization, enabling improved risk stratification and candidate selection for active surveillance (AS). However, little is known about the clinical outcomes associated with small mpMRI lesions, which may be associated with low tumor burden and a potentially less aggressive course. This study aims to define the rates of prostate cancer progression and radical therapy for patients on AS with small index lesions and to explore the relationship between index lesion growth rates and cancer progression. A prospectively maintained database of patients on active surveillance at the National Cancer Institute was queried for patients with mpMRI index lesions ≤7mm in greatest diameter who had ≥2 mpMRIs and MRI-Ultrasound fusion biopsy sessions. Patients were identified with mpMRI lesions confirmed on MRI/Ultrasound fusion biopsy to have no worse than Gleason grade group (GG) 1 prostatic adenocarcinoma. Index lesions were defined as the MRI lesions with the highest PI-RADS score, with size used as a tiebreaker. For MRIs preceding PI-RADS scoring, an internal Likert suspicion scoring system that has been validated in the literature was used and converted into PI-RADS scores. MR review was conducted by a high-volume dedicated genitourinary radiologist.;Time to GG progression was calculated to be the difference between initial mpMRI and mpMRI at the time of cancer progression. Between 2007-2022, patients on AS with small index lesions measuring ≤7 mm (n=88) were identified. 26.53% of patients whose index lesions which grew between 0 and 1 mm/year (13/49) progressed to GG2 disease, with an average time to progression of 5.28 years. Of these patients, 12.24% (6/49) progressed to radical therapy, on average 6.16 years after initiating AS (Table 1). Eight years after initial mpMRI, 28.41% of patients in the full cohort (25/88) progressed to radical therapy (Figure 1). Overall, only 5.68% (5/88) of patients developed GG3 disease within 2 years of follow-up, and 2.27% (2/88) progressed to radical therapy. Our findings have important implications for timing of follow-up for patients on AS with small index lesions. The low rate of clinical progression to GG2 disease and radical therapy indicates a less aggressive clinical course than prior published AS series in the pre-MRI era. The presence of small index lesions;may therefore be considered as a favorable marker for determining eligibility and risk of progression for patients considering AS and may guide the timing of repeat imaging to personalize surveillance strategies. [ABSTRACT FROM AUTHOR]

To evaluate CycleGAN's ability to enhance T2-weighted image (T2WI) quality. A CycleGAN algorithm was used to enhance T2WI quality. 96 patients (192 scans) were identified from patients who underwent multiple axial T2WI due to poor quality on the first attempt (RAD1) and improved quality on re-acquisition (RAD2). CycleGAN algorithm gave DL classifier scores (0–1) for quality quantification and produced enhanced versions of QI1 and QI2 from RAD1 and RAD2, respectively. A subset (n = 20 patients) was selected for a blinded, multi-reader study, where four radiologists rated T2WI on a scale of 1–4 for quality. The multi-reader study presented readers with 60 image pairs (RAD1 vs RAD2, RAD1 vs QI1, and RAD2 vs QI2), allowing for selecting sequence preferences and quantifying the quality changes. The DL classifier correctly discerned 71.9 % of quality classes, with 90.6 % (96/106) as poor quality and 48.8 % (42/86) as diagnostic in original sequences (RAD1, RAD2). CycleGAN images (QI1, QI2) demonstrated quantitative improvements, with consistently higher DL classifier scores than original scans (p < 0.001). In the multi-reader analysis, CycleGAN demonstrated no qualitative improvements, with diminished overall quality and motion in QI2 in most patients compared to RAD2, with noise levels remaining similar (8/20). No readers preferred QI2 to RAD2 for diagnosis. Despite quantitative enhancements with CycleGAN, there was no qualitative boost in T2WI diagnostic quality, noise, or motion. Expert radiologists didn't favor CycleGAN images over standard scans, highlighting the divide between quantitative and qualitative metrics. [ABSTRACT FROM AUTHOR]