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.

A new tumorgraft panel to accelerate precision medicine in prostate cancer.

Background: Despite the significant advances in the management of advanced prostate cancer (PCa), metastatic PCa is currently considered incurable. For further investigations in precision treatment, the development of preclinical models representing the complex prostate tumor heterogeneity are mandatory. Accordingly, we aimed to establish a resource of patient-derived xenograft (PDX) models that exemplify each phase of this multistage disease for accurate and rapid evaluation of candidate therapies. Methods: Fresh tumor samples along with normal corresponding tissues were obtained directly from patients at surgery. To ensure that the established models reproduce the main features of patient's tumor, both PDX tumors at multiple passages and patient's primary tumors, were processed for histological characteristics. STR profile analyses were also performed to confirm patient identity. Finally, the responses of the PDX models to androgen deprivation, PARP inhibitors and chemotherapy were also evaluated. Results: In this study, we described the development and characterization of 5 new PDX models of PCa. Within this collection, hormone-naïve, androgen-sensitive and castration-resistant (CRPC) primary tumors as well as prostate carcinoma with neuroendocrine differentiation (CRPC-NE) were represented. Interestingly, the comprehensive genomic characterization of the models identified recurrent cancer driver alterations in androgen signaling, DNA repair and PI3K, among others. Results were supported by expression patterns highlighting new potential targets among gene drivers and the metabolic pathway. In addition, in vivo results showed heterogeneity of response to androgen deprivation and chemotherapy, like the responses of patients to these treatments. Importantly, the neuroendocrine model has been shown to be responsive to PARP inhibitor. Conclusion: We have developed a biobank of 5 PDX models from hormone-naïve, androgen-sensitive to CRPC primary tumors and CRPC-NE. Increased copy-number alterations and accumulation of mutations within cancer driver genes as well as the metabolism shift are consistent with the increased resistance mechanisms to treatment. The pharmacological characterization suggested that the CRPC-NE could benefit from the PARP inhibitor treatment. Given the difficulties in developing such models, this relevant panel of PDX models of PCa will provide the scientific community with an additional resource for the further development of PDAC research.

Radical Prostatectomy after Vascular Targeted Photodynamic Therapy with Padeliporfin: Feasibility, and Early and Intermediate Results.

PURPOSE: Vascular targeted photodynamic therapy with TOOKAD® is a new therapeutic option for localized prostate cancer management. The objectives of this study were to assess the feasibility of radical prostatectomy after vascular targeted photodynamic therapy and describe functional and oncologic outcomes. MATERIALS AND METHODS: We retrospectively included in study 45 patients who underwent salvage radical prostatectomy after vascular targeted photodynamic therapy for recurrent prostate cancer at a total of 14 surgical centers in Europe between October 2008 and March 2017. Of the 42 radical prostatectomies performed 16 were robot-assisted, 6 were laparoscopic and 20 were open surgery. Primary end points were morbidity and technical difficulties. Secondary end points were early and intermediate postoperative functional and oncologic outcomes. RESULTS: Median operative time was 180 minutes (IQR 150-223). Median blood loss was 200 ml (IQR 155-363). According to the surgeons the surgery was easy in 29 patients (69%) and difficult in 13 (31%). Nerve sparing was feasible in 14 patients (33%). Five postoperative complications (12%) were found, including 2 Clavien I, 2 Clavien II and 1 Clavien IIIB complications. Of the cases 13 (31%) were pT3 and 21 (50%) were pT2c. Surgical margins were positive in 13 patients (31%). Prostate specific antigen was undetectable at 6 to 12 months in 37 patients (88%). Nine patients underwent complementary radiotherapy. Four patients had final prostate specific antigen greater than 0.2 ng/ml at a median followup of 23 months (IQR 12-36). At 1 year 27 patients (64%) were completely continent (no pads) and 10 (24%) had low incontinence (1 pad). Four patients (11%) recovered potency without treatment and 23 (64%) recovered potency with appropriate treatment. CONCLUSIONS: Salvage radical prostatectomy after vascular targeted photodynamic therapy treatment was feasible and safe without difficulty for most of the surgeons.

Laparoscopic sentinel lymph node dissection in prostate cancer patients: the additional value depends on preoperative data.

AIM: In intermediate- or high-risk prostate cancer (PC) patients, to avoid extended pelvic lymph node dissection (ePLND), the updated Briganti nomogram is recommended with the cost of missing 1.5 % of patients with lymph node invasion (LNI). Is it possible to reduce the percentage of unexpected LNI patients (nomogram false negative)? We used the isotopic sentinel lymph node (SLN) technique systematically associated with laparoscopic ePLND to assess the potential value of isotopic SLN method to adress this point. METHODS: Two hundred and two consecutive patients had procedures with isotopic SLN detection associated with laparoscopic ePLND for high or intermediate risk of PC. The area under the curve (AUC) of the receiver operating characteristics (ROC) analysis was used to quantify the accuracy of different models as: the updated Briganti nomogram, the percentage of positive cores, and an equation of the best predictors of LNI. We tested the model cutoffs associated with an optimal negative predictive value (NPV) and the best cutoff associated with avoiding false negative SLN detection, in order to assist the clinician's decision of when to spare ePLND. RESULTS: LNI was detected in 35 patients (17.2 %). Based on preoperative primary Gleason grade and percentage of positive cores, a bivariate model was built to calculate a combined score reflecting the risk of LNI. For the Briganti nomogram, the 5 % probability cutoff avoided ePLND in 53 % (108/202) of patients, missing three LNI patients (8.6 %), but all were detected by the SLN technique. For our bivariate model, the best cutoff was <10, leaving no patient with LNI due to positive SLN detection (four patients = 11.4 %), and avoiding ePLND in 52 % (105/202) of patients. CONCLUSION: For patients with a low risk of LNI determined using the updated Briganti nomogram or bivariate model, SLN technique could be used alone for lymph node staging in intermediate- or high-risk PC patients.

Laparoscopic sentinel lymph node versus hyperextensive pelvic dissection for staging clinically localized prostate carcinoma: a prospective study of 200 patients.

UNLABELLED: Lymph node metastasis is an important prognostic factor in prostate cancer (PC). The aim of this prospective study was to validate, through laparoscopic surgery, the accuracy of the isotopic sentinel lymph node (SLN) technique correlated with hyperextensive pelvic resection (extended pelvic lymphadenectomy dissection) in patients with localized PC, candidates for local curative treatment. METHODS: A transrectal ultrasound-guided injection of (99m)Tc-sulfur rhenium colloid (0.3 mL/100 MBq) in each prostatic lobe was performed the day before surgery. Detection was performed intraoperatively with a laparoscopic probe, followed by extensive resection. SLN counts were performed in vivo and confirmed ex vivo. Histologic analysis was performed by hematoxylin-phloxine-safran staining, followed by immunohistochemistry if the SLN was free of metastasis. RESULTS: Two hundred three patients with PC at intermediate or high risk of lymph node metastases were included. The intraoperative detection rate was 96% (195/203). Thirty-five patients had lymph node metastases, 19 only in the SLN. The false-negative rate was 8.5% (3/35). Unilateral surgical SLN detection did not validate bilateral pelvic lymph node status, and extended pelvic lymphadenectomy dissection was necessary on the opposite side of detection to minimize the false-negative rate (2.8% [1/35]). A significant metastatic sentinel invasion in the common iliac region existed (9.3%) but was always associated with other metastatic node areas. The internal iliac region was the primary metastatic site (40.7%). Finally, this series invalidated any justification for a standard or limited dissection, which would have missed 51.9% and 74.1% of lymph node metastases, respectively. CONCLUSION: The radioisotope SLN identification method up to the common iliac region is successful to identify sentinel nodes during laparoscopic surgery per hemipelvis to be acceptably considered as an isolated procedure and should be validated for intermediate- and high-risk patients.

Dynamic contrast-enhanced-magnetic resonance imaging evaluation of intraprostatic prostate cancer: correlation with radical prostatectomy specimens.

OBJECTIVES: To determine the diagnostic performance of dynamic contrast-enhanced-magnetic resonance imaging (DCE-MRI) in the identification of intraprostatic cancer foci related to cancer volume at histopathology, in patients with clinically localized cancer treated by radical prostatectomy, with whole-mount histopathologic sections as the reference standard. METHODS: Eighty-three consecutive radical prostatectomy specimens from patients referred for a prostate-specific antigen elevation were correlated with prebiopsy MRI. MRI results ranked on a 5-point scale were correlated with the findings of histopathology maps in 8 prostate sectors, including volume, largest surface area, and percentage of Gleason grade 4/5. The area under the receiver operating characteristic curve was used. RESULTS: Median prostate-specific antigen was 8.15 ng/mL. DCE-MRI was suspicious in 55 (66%) out of 83 patients. A separate cancer foci (mean 2.55 per patient) was present in 212 (34%) of 664 octants and DCE-MRI was suspicious in 68 of 212. Sensitivity and specificity of DCE-MRI at score 3.4 or 5 for identification of cancer foci at any volume was 32% and 95%, respectively. For identification of cancer foci > 0.5 mL, the sensitivity and specificity were 86% and 94%, respectively, with the under the receiver operating characteristic curve of 0.874. Mean volume of DCE-MRI detected and missed cancers were 2.44 mL (0.02-14.5) and 0.16 mL (0.005-2.4), respectively. Sensitivity and specificity of DCE-MRI for identification of > 10% of Gleason grade 4/5 were 81% and 82%, respectively. CONCLUSIONS: DCE-MRI can accurately identify intraprostatic cancer foci. Possible applications are guidance for biopsies, selection of patients for watchful waiting, and focal treatment planning.

Transition zone and anterior stromal prostate cancers: zone of origin and intraprostatic patterns of spread at histopathology.

BACKGROUND: To describe the precise location of transition zone (TZ) and anterior fibromuscular stroma (AFMS) prostate cancers (TZ/AFMS) within histological zones at various stages of development and to demonstrate their pattern of intraprostatic spread from their site of origin. METHODS: Anterior TZ/AFMS cancers excluding the anterolateral part of peripheral zone, were identified from radical prostatectomy specimens. Morphometric histopathological study included largest surface area, volume and spatial distribution. RESULTS: Out of 91 TZ/AFMS cancers, 79 were <4 cm3 and 69 <2 cm3. Fifty percent and 70% of cancers <4 cm3 were located in the anterior third and inferior half of TZ and/or AFMS, respectively. Cancers <2 cm3 could be classified into three types according to their location related to histologic zone boundaries: TZ type 1 (40%) for cancers confined to one TZ lobe; TZ type 2 (35%) for cancers most represented in one TZ lobe but crossing its anterior boundary; type AFMS (25%) for cancers confined to AFMS. These results form the rationale for the hypothesis that AFMS cancers originate from anterior and medial TZ and due to benign prostatic hypertrophy they become excluded from TZ, anteriorly into AFMS. TZ anterior limit would then act as a barrier to their posterior extension. CONCLUSIONS: TZ/AFMS cancers contours and locations are predictable and conform to histological zones boundaries. Knowledge of these cancer origin and pattern of spread in TZ and AFMS are of importance for imaging diagnosis, guidance for biopsy and focal therapy.

Peripheral zone prostate cancers: location and intraprostatic patterns of spread at histopathology.

BACKGROUND: To describe the precise location of peripheral zone (PZ) prostate cancers at various stages of development and to demonstrate their pattern of intraprostatic spread from their site of origin. METHODS: PZ cancers defined as cancers located in peripheral zone (PZ) including the anterolateral part of PZ, were identified from radical prostatectomy specimens. PZ cancers morphometric histopathological study included largest surface area, volume and spatial distribution. RESULTS: Out of 188 PZ cancers, 179 were <4 cm(3) and 168 <2 cm(3). PZ cancers were still confined to their zone of origin for volumes <2 cm(3). Between 2 and 4 cm(3), some cancers partially spread into the transition zone or anterior fibromuscular stroma. Sixty-four and 90% of PZ cancers <4 cm(3) were located in the lower and posterior half of the gland respectively. Ten percent were located in the anterior horn of PZ. Overall, non-index (second) cancers were located in the ipsi and contolateral side of the index cancer (largest) in 31% and 69% of cases, respectively. Cancers <2 cm(3) were confined to one lobe in 164 of 168 (98%) cases and not confined in 3 out of 11 (27%) cancers 2-4 cm(3). On vertical axis, only cancers >or=2 cm(3) involved both apex and base. CONCLUSIONS: PZ cancers contours and locations are predictable and conform to histological zone boundaries if <2 cm(3) in volume. Knowledge of PZ cancers origin and pattern of spread in PZ are of importance for imaging diagnosis, guidance for biopsy and focal therapy.