Histological differences in cancer cells, stroma, and luminal spaces strongly correlate with in vivo MRI-detectability of prostate cancer

• Published in: Modern pathology Vol. 32; no. 10; pp. 1536 - 1543
• Main Authors: Miyai, Kosuke, Mikoshi, Ayako, Hamabe, Fumiko, Nakanishi, Kuniaki, Ito, Keiichi, Tsuda, Hitoshi, Shinmoto, Hiroshi
• Format: Journal Article
• Language: English
• Published: New York Nature Publishing Group US 01.10.2019; Elsevier Limited
• Subjects: 13/51; 59/57; 631/67/2321; 631/67/589/466; Aged; Cancer surgery; Humans; Image Interpretation, Computer-Assisted - methods; Laboratory Medicine; Magnetic resonance imaging; Male; Medicine; Medicine & Public Health; Middle Aged; Multiparametric Magnetic Resonance Imaging - methods; NMR; Nuclear magnetic resonance; Pathology; Prostate cancer; Prostatectomy; Prostatic Neoplasms - diagnostic imaging; Prostatic Neoplasms - pathology; Retrospective Studies; Stroma; Urological surgery
• ISSN: 0893-3952; 1530-0285
• DOI: 10.1038/s41379-019-0292-y

The current study aimed to investigate the plausible histopathological factors that affect the detectability of prostate cancers on multiparametric magnetic resonance imaging (MP-MRI). This retrospective study included 59 consecutive patients who had undergone MP-MRI and subsequent radical prostatectomy. The cases were standardized according to the tumor size ranging from 10 to 20 mm on the final pathological diagnosis. Histopathological review and semi-automated imaging analysis were performed to evaluate the relative area fractions of the histological components, including cancer cells, stroma, and luminal spaces. Among the 59 prostatectomy specimens, no case showed two or more foci of cancer that matched the size criteria. Of the 59 lesions, 35 were MRI-detectable [Prostate Imaging Reporting and Data System (PIRADS) score of 3 or greater] and 24 were MRI-undetectable (PIRADS score of 2 or less). No significant differences were observed in Gleason Grade Group, percentage of Gleason pattern 4, and predominant subtype of Gleason pattern 4 between MRI-detectable and MRI-undetectable cancers. On the other hand, significantly higher mean area fraction of cancer cells (60.9% vs. 42.7%, P  < 0.0001) and lower mean area fractions of stroma (33.8% vs. 45.1%, P  = 0.00089) and luminal spaces (5.2% vs. 12.2%, P  < 0.0001) were observed in MRI-detectable cancers than in MRI-undetectable cancers. In a multivariable analysis performed upon exclusion of area fraction of stroma due to its multicollinearity with that of cancer cells, area fractions of cancer cells ( P  = 0.0031) and luminal space ( P  = 0.0035) demonstrated strong positive and negative correlation with MRI-detectability, respectively. Changes in cancer cells, stroma, and luminal spaces, rather than conventional histological parameters, could be considered one of the best predictors to clinical, in vivo MRI-detectability of prostate cancer.