Prostate Cancer Grading: A Decade After the 2005 Modified Gleason Grading System Open Access

We wish to comment on the prostate cancer grading review by Kryvenko and Epstein,¹  published in Archives of Pathology & Laboratory Medicine. First, we would point out that, contrary to what was stated in the review, the first publications on prostate cancer predate those of Hugh Hampton Young. In his classic text De Sedibus et Causis Morborum, published in 1761, Morgagni refers to 2 cases that were almost certainly carcinoma. Two cases of carcinoma were reported by Brodie in 1832, and, indeed, many other series were published before 1900, including the definitive study of Thompson that showed the tumor to occur commonly.² 

In their review, the authors report that the International Society of Urological Pathology (ISUP) grading system was validated by a study of 20 845 patients from 5 institutions.³  In reality, these data do not validate the ISUP grading system because cases were accessioned from 2005 to 2014. The 2005 ISUP consensus conference was held in March 2005, with the definitive publication appearing in September 2005. Although we can accept that the Johns Hopkins Hospital (JHH) material may have been graded according to 2005 criteria, the same does not necessarily apply to the other 4 institutions. Indeed, it is the experience of one of us who contributed to that study, that the 2005 grading criteria took some years to be fully embraced locally. In view of this, the lack of a central review means that data from the 2014 “validation” study must be considered with some caution. Although these have not been referenced by the authors of the review, validation studies of ISUP grading have been reported.^4,5  In these 2 studies, ISUP grade correlated with distant progression-free and prostatic-specific antigen progression-free survival. Despite the significant predictive value of ISUP grading, there is considerable overlap in outcome among some grades. This especially involves ISUP grades 1 and 2, as well as grades 3 and 4. The ISUP grades 1 and 2 overlap is probably due to ISUP grade 2 tumors with limited volume pattern 4 having an outcome similar to ISUP grade 1 (3 + 3 = 6) tumors. The overlap in outcome for ISUP grades 3 and 4 tumors is likely the result of lumping 3 + 5, 5 + 3, and 4 + 4 tumors into ISUP grade 4 (as discussed below). This suggests that the current grading system could be reconfigured to a 4- or even a 3-tier system.

Contrary to the claims of the authors, the consensus conference did not resolve the issues regarding the handling of tertiary grades in the reporting of radical prostatectomy specimens. They stated that tertiary patterns of more than 5% of the tumor nodule should be incorporated into the Gleason score as the secondary pattern (regardless of the volume of the true secondary pattern tumor). At the 2005 ISUP conference, it was agreed that this handling would apply to needle biopsies only. Radical prostatectomy specimens were not similarly upgraded because it was considered that, using the Gleason classic scoring, needle-biopsy scores underestimated the final Gleason score obtained from examination of the radical prostatectomy specimen. The recommendation of the 2005 conference relating to the methodology of deriving a Gleason scoring for radical prostatectomy specimens was not altered at the 2014 consensus conference.^6,7  As a consequence, the suggestion that 3 + 4 = 7 (tertiary 5) in radical prostatectomy specimens be upgraded to 3 + 5 = 8 (ISUP grade 4) is not in accordance with the conference recommendations. In effect, the 2014 ISUP grading cannot be applied to radical prostatectomy specimens because there is no provision for the accommodation of tertiary scores.

In the review, the senior author makes a strong claim to be the originator of the 2014 ISUP grading system for prostate cancer. What he failed to mention is that there was a group of 6 pathologists appointed by ISUP Council to coordinate the conference, and these pathologists were the principal authors of the definitive publication.⁷  An early decision of the group was that the grading classification would be known as ISUP grading. The matter was later addressed by the ISUP Council, and it was unanimously resolved that, in parallel with all other grading and classification systems developed under the auspices of the ISUP,^6,8–10  the term ISUP grade would be applied. The term grade grouping is unfortunate because it is inaccurate as groupings are based on score as well as grade, rather than being based on grade alone. Further, ISUP grading is not simply a grouping of grades/scores but includes revised morphologic criteria generated by participants at both the 2005 and 2014 consensus conferences of the ISUP.^6,7 

The authors state the new grading system was first described in work undertaken at JHH; however, we would take issue with that. The concept of grouping components of the Gleason system is hardly novel and, in fact, was proposed by Gleason himself in 1977. As we have noted elsewhere,¹¹  a variety of groupings have been previously suggested, with the grouping proposed by Donohue et al¹²  being almost identical to those in the Pierorazio 2013 study.¹³ 

In reality, ISUP grade differs from JHH 2013 grades because JHH grade 4 consists of 4 + 4 = 8 tumors only, whereas ISUP grade 4 also includes 5 + 3 = 8 and 3 + 5 = 8 tumors.^7,13  Perhaps more important is the following question: Are tumors with these scores likely to behave in a manner similar to 4 + 4 = 8 tumors? Growing evidence would indicate that 5 + 3 = 8 tumors are more appropriately classified as ISUP grade 5, and 2 recent studies have validated this suggestion.¹¹  The authors of the review mentioned those studies and attempted to dismiss those important findings with the suggestion that the grading criteria have changed. This same comment can be applied to the Epstein et al³ “validation” study that occurred in 2014, which, as we have noted, contains cases accessioned before the publication of the 2005 ISUP modified grading criteria. The reviewers also note that the 5 + 3/3 + 5 tumors are uncommon; however, this does not mean that patients with prostate cancer should not have their tumors graded and treated appropriately. Further, the true incidence of 3 + 5/5 + 3 tumors depends on whether grading is based on a single core or multiple core (case-composite) approach. In view of this evidence, it would appear that ISUP grading categories will require revision in the immediate future.