Our study demonstrated that prostate volume and biopsy tumor length had independent value for predicting both classical and redefined insignificant cancer, and PSAD showed the independent value for only the redefined insignificant cancer. The different statistics between classical and redefined insignificant cancers in the multivariate analysis might imply that PSAD possibly holds predictive power in larger or aggressive tumors. Substantial overlaps existed in cases with small prostate volume, but large prostate volume firmly had high positive predictive value for tumors of stage ≤ pT2 and negative surgical margins, namely organ-confined cancers. We consider that the finding that prostate volume is important does not simply mirror many other studies showing that PSAD is important. Furthermore, clinical T-staging had little value, and the multiparametric MRI would possibly add some diagnostic value if evaluation was performed in detail.
The favorable features on multiple core biopsies consistently harbor unfavorable pathologic results on prostatectomy specimens despite fulfilling the established criteria developed in North America and Europe. Certainly, in our data, patients meeting Epstein and PRIAS criteria harbored clinically significant cancer for one-third and around half of cases, respectively. Oon et al. speculated that the modification to Gleason scoring might be associated with a reduced accuracy of Epstein criteria, because distinct differences were observed in validation studies between pre- and post-2005
[11, 18]. In addition, Wolters, et al. presented a recent analysis using a data set from a randomized screening trial that demonstrated that clinically insignificant prostate cancer may include GS 6, pT2 tumors with index tumor volumes of up to 1.3 mL
 instead of 0.5 mL, which had been used as a threshold for around 20 years
. These critical alterations suggest reconsideration of the current methods used for risk assessment before definitive therapy or AS. The established criteria consisting of clinical T-staging by digital rectal examination or PSAD would not be satisfactory for patients in this study, and the clinical criteria should be compiled incorporating the prostate size and biopsy length involved by cancer.
Prostate volume has been less often mentioned than PSAD or PSA as a predictive factor despite that all of these parameters held independent values
. PSAD is a comprehensive parameter considering both serum PSA and prostate size, but the value varies easily based on the fluctuation of PSA. In that sense, prostate volume estimated using MRI is a stable preoperative parameter and the only tumor-unrelated factor in our analysis model. Despite the variations of study design and whether the prostate volume was estimated pre- or postoperatively, small prostate volume is unanimously reported with association between the poor oncological outcomes in several studies. Freedland et al. reported that more high-grade and more advanced cancers were detected in men with smaller prostates along with lower serum testosterone concentrations in a large population ranging from clinical T1 to T3 cancers and suggested that prostate size might be an important prognostic variable that should be evaluated for use pre- and postoperatively
. Tilki et al. and Chung et al. reported the associations especially in relation to GS upgrading
[21, 23]. These trends were also observed even when study populations were limited with GS ≤ 6 by Milonas et al.
 and with highly selective criteria (T1c, PSA < 10 ng/mL, a single positive biopsy, tumor length < 3 mm, and Gleason score < 7) by Beauval et al.
. The prostate volume would be directly affected by age and endocrine factors
, and the mean prostate size should be significantly different between Japanese and Western populations even after adjusting for differences in age, height, and weight
. Although the median prostate volume was obviously different between 35.5 cc in our study and more than 40 to 50 cc in the Western study, our study confirmed that prostate volume retained its predictive ability for both classical and redefined insignificant cancer in Japanese men.
The percentage and length of cancer involvement in biopsy core are also significant predictors, and have already been incorporated in the major prediction models before developing AS protocols
. Russo et al. reported that inclusion of the percentage of cancer involvement contributed to reducing the misclassification in patients eligible for AS according to the PRIAS criteria, which does not reference any cancer involvement in the core
. Antonelli et al. used the updated definition of total tumor volume and determined its optimal cutoff to be 20% for the diagnosis of insignificant cancer using the receiver operating characteristic curve
. Freedland et al. reported the same threshold of 20% for predicting PSA recurrence after prostatectomy
. We agree with their strict threshold despite the fact that our data were analyzed in terms of tumor length but not in the percentage of tumor involvement, and we consider the threshold of 50%, which is used in the many AS protocols, to be too relaxed to avoid under-estimation of cancer. However, a more stringent threshold, namely minute or microfocal cancer defined by ≤ 5% or ≤ 1 mm in a single core, is not a guarantee of insignificant cancer
The current established clinical criteria can-not eliminate the risk of over- and under- estimation of cancer. A stringent selection criteria excludes a considerable number of patients who are willing to be managed by AS, even those having potentially insignificant cancer, and therefore, well-balanced criteria between sensitivity and specificity are required for patients. The implication of increasing the index tumor volume threshold to 1.3 mL is that we should miss small-volume cancer and set the AS protocols to be more expanded. In the recent report of head-to-head comparison of contemporary AS protocols, Iremashvili et al. revealed that the PRIAS and University of Miami criteria demonstrated the best balance between sensitivity and specificity for insignificant prostate cancer among the existing AS protocols, and the contemporary Epstein criteria demonstrated high specificity but low sensitivity for all end points
Nevertheless, clinically significant cancer might not always progress, and some cases might remain indolent for a substantial duration of time. A validation of AS protocol should not be a surrogate endpoint, such as the analysis of the pathologic results of prostatectomy specimens, but should be a long-term outcome of a prospective cohort. According to the review of AS in the large prospective series, approximately one-third of patients were treated after a median surveillance of about 2.5 years because of histologic reclassification on biopsy or a PSA doubling time of less than 3 years, while some cases were treated with no evidence of progression
. The short- to mid-term estimated treatment-free survivals were reported as 62 to 72% at 5 years and at 43 to 62% at 10 years
[34, 35]. These data characterize AS as a strategy for deferring treatment and justify it as an optimal choice for patients with low-risk PCa that can accept the confirmatory biopsy within 1 to 2 years and the slight increased risk of late metastasis.
To develop a model to discriminate clinically indolent from aggressive disease efficiently, advances in biochemical markers replacing PSA or PSA derivatives such as prostate cancer antigen 3 or transmembrane protease serine 2 will be required in addition to the existing factors
. In addition, a more detailed analysis of multiparametric MRI, including number of lesions, lesion suspicion, and lesion density (calculated as total lesion volume/prostate volume)
, and image-guided targeted biopsy should play a positive role
. The current study has some limitations; there is no control population outside of Japan other than the published literature. It is a retrospective study based on a relatively small Japanese population, and the pathologic examination was performed at 2 institutions. The median follow-up time was also relatively short to determine oncological outcomes. Thus, the results may not apply to the Western population. Nonetheless, it is valuable to give insight into ethnic differences, and these data provide useful information that could help predict insignificant cancer in Japanese or Asian patients with favorable features on needle biopsy. The findings of this study should be validated in a larger, independent dataset.