Faculty Bibliography

Accurate recognition of Gleason pattern (GP) 4 prostate carcinoma (PCa) on needle biopsy is critical for patient management and prognostication. "Poorly formed glands" are the most common GP4 subpattern. We studied the diagnostic reproducibility and the quantitative threshold of grading GP4 "poorly formed glands" and the criteria to distinguish them from tangentially sectioned GP3 glands. Seventeen urologic pathologists were first queried for the definition of "poorly formed glands" using cases representing a spectrum of PCa glandular differentiation. Cancer glands with no or rare lumens, elongated compressed glands, and elongated nests were considered "poorly formed glands" by consensus. Participants then graded a second set of 23 PCa cases that potentially contained "poorly formed glands" with a fair interobserver agreement (kappa=0.34). The consensus diagnoses, defined as agreement by >70% participants, were then correlated with the quantitative (10) and topographic features of poorly formed glands (clustered, immediately adjacent to, and intermixed with other well-formed PCa glands) in each case. Poorly formed glands immediately adjacent to other well-formed glands regardless of their number and small foci of 10 poorly formed glands that were not immediately adjacent to well-formed glands were graded as GP4. Grading "poorly formed glands" is challenging. Some morphologic features are, however, reproducible for and against a GP4 diagnosis. This study represents an important step in standardization of grading of "poorly formed glands" based on quantitative and topographic morphologic features.

Accurate recognition of Gleason pattern 5 (GP5) prostate adenocarcinoma on needle biopsy is critical as it is associated with disease progression and adverse clinical outcome. Despite important implications of this diagnosis, interobserver variation in the diagnosis of GP5 has not been adequately studied. Digital images of 66 prostate adenocarcinoma cases that potentially contained a GP5 component were distributed to 16 urologic pathologists who were asked to classify whether GP5 was present. Each image was initially classified into 1 of 4 morphologic subpatterns by 2 coauthors (R.B.S. and M.Z.): solid nests (15), comedocarcinoma (8), single cells and/or cords (35), and variant morphology (8). Additional features captured included: size (large: >20 cells, medium: 10 to 20 cells, and small: <10 cells) and distribution of nuclei (uniform vs. nonuniform) for nests pattern; intraluminal coagulative tumor necrosis, karyorrhectic debris, and amorphous material for comedocarcinoma pattern; and quantity (10) and distribution (clustered vs. intermixed with adjacent well-formed glands) for single cells/cords pattern. Interobserver reproducibility of a diagnosis of GP5 was assessed and the morphologic subpatterns and features were correlated with the consensus diagnosis (defined as 75% agreement). Interobserver reproducibility for overall diagnostic agreement was fair (kappa=0.376). Among subpatterns, comedocarcinoma had highest reproducibility (kappa=0.499), followed by variant morphology (kappa=0.443), single cells/cords (kappa=0.369), and nests (kappa=0.347). All cases with the following features achieved consensus for GP5: large nests regardless of nuclear distribution; coagulative necrosis with or without karyorrhectic debris; single cells/cords >10 or 6 to 10 in a cluster; and signet ring-like cells in single cells or within nests pattern. A majority of cases with the following features achieved consensus against GP5: medium-size nests; exclusive intraluminal amorphous material; single cells/cords 1 level. The diagnostic reproducibility of GP5 within certain morphologies was only fair among urologic pathologists. However, the diagnosis of GP5 was more reproducible when certain restrictive morphologic and quantitative criteria were applied. These findings suggest that additional studies are needed to find highly reproducible features of GP5 associated with documented aggressive clinical outcome.

PURPOSE: To compare boundaries of prostate tumors on MRI and histologic assessment from radical prostatectomy (RP) using detailed software-assisted co-registration, in order to define an optimal treatment margin to achieve complete tumor destruction during image-guided focal ablation. METHODS: 33 patients who underwent 3T MRI before RP were included. A radiologist traced lesion borders on MRI and assigned a suspicion score (SS) from 2-5. 3D reconstructions were created from high-resolution digitalized slides from RP specimens and co-registered to MRI using advanced software. Tumors were compared between histology and MRI using the Hausdorff Distance (HD) and stratified by MRI-SS, Gleason Score (GS), and lesion diameter. Cylindrical volume estimates of treatment effects were used to define the optimal treatment margin. RESULTS: 46 histologically confirmed cancers underwent 3D software-based registration with MRI. MRI underestimated tumor sizes, with the maximal discrepancy between MRI and histologic boundaries for a given tumor averaging 1.99+/-3.1mm (18.5% of the MRI diameter). Boundary underestimation was larger for MRI-SS>/=4 lesions (+3.49+/-2.1mm; p<0.001) and GS>/=7 lesions (+2.48+/-2.8mm; p 0.035). On average, a simulated cylindrical treatment volume based on the MRI boundary missed 14.8% of the tumor volume compared with a simulated cylindrical volume based on the histologic boundary. A simulated treatment volume based on a 9mm treatment margin achieved complete histologic tumor destruction in 100% of patients. CONCLUSION: MRI underestimates histologically-determined tumor boundaries, especially for high MRI-SS and high GS lesions. A 9mm treatment margin around an MRI-visible lesion consistently ensures treatment of the entire histologic tumor volume during focal ablative therapy.

PURPOSE: To retrospectively assess the utility of whole-lesion apparent diffusion coefficient (ADC) metrics in characterizing the Gleason 4 component of Gleason 7 prostate cancer (PCa) at radical prostatectomy. MATERIALS AND METHODS: Seventy patients underwent phased-array coil 3T-magnetic resonance imaging (MRI) before prostatectomy. A uropathologist mapped locations and Gleason 4 percentage (G4%) of Gleason 7 tumors. Two radiologists independently reviewed ADC maps, aware of tumor locations but not G4%, and placed a volume-of-interest (VOI) on all slices including each lesion on the ADC map to obtain whole-lesion mean ADC and ADC entropy. Entropy reflects textural variation and increases with greater macroscopic heterogeneity. Performance for characterizing Gleason 7 tumors was assessed with mixed-model analysis of variance (ANOVA) and logistic regression. RESULTS: Among 84 Gleason 7 tumors (G4% 5%-85%, median 30%; 59 Gleason 3+4, 25 Gleason 4+3), ADC entropy was significantly higher in Gleason 4+3 than Gleason 3+4 tumors (R1: 5.27 +/- 0.61 vs. 4.62 +/- 0.78, P = 0.001; R2: 5.91 +/- 0.32 vs. 5.57 +/- 0.56, P = 0.004); mean ADC was not significantly different between these groups (R1: 0.90 +/- 0.15*10-3 cm2 /s vs. 0.98 +/- 0.21*10-3 cm2 /s, P = 0.075; R2: 1.06 +/- 0.19*10-3 cm2 /s vs. 1.14 +/- 0.16*10-3 cm2 /s, P = 0.083). The area under the receiver operating characteristic (ROC) curve (AUC) for differentiating groups was significantly higher with ADC entropy than mean ADC for one observer (R1: 0.74 vs. 0.57, P = 0.027; R2: 0.69 vs. 0.61, P = 0.329). For R1, correlation with G4% was moderate for ADC entropy (r = 0.45) and weak for mean ADC (r = -0.25). For R2, correlation with G4% was moderate for ADC entropy (r = 0.41) and mean ADC (r = -0.32). For both readers, ADC entropy (P = 0.028-0.003), but not mean ADC (P = 0.384-0.854), was a significant independent predictor of G4%. CONCLUSION: Whole-lesion ADC entropy outperformed mean ADC in characterizing Gleason 7 tumors and may help refine prognosis for this heterogeneous PCa subset. J. Magn. Reson. Imaging 2014. (c) 2014 Wiley Periodicals, Inc.