Post-baseline biopsies

1. It would be best to report out two summaries for each pathological endpoint. One would include only the Central review data (i.e., data from Bostwick Labs contained in the PATHOLOG SAS dataset), and the other would include Central lab and local lab data (i.e., data from the PATHOLOG and PATHO datasets, with PATHO data only used for biopsies in which no Central lab data for that subject are available) (Matt's e-mail: Fri, 12 Aug 2005 11:45:02 -0400).

Unscheduled prostate biopsies

1. Cancer, ASAP (atypical small proliferation), Suspicious.
2. HGPIN (high grade prostatic intraepithelial neoplasia)
3. HGPIN volume volume of HGPIN - should be present in the database for each post-baseline core that has a HGPIN diagnosis; cores with no HGPIN diagnosis will need to be set to zero. Then sum all HGPIN volume values across a given biopsy. This allows summarization for both all subjects and for a "HGPIN" population (Matt's e-mail: Tue, 16 Aug 2005 09:56:58 -0400).
4. Gleason score
5. % core involved and
6. # of cores with cancer - pertinent values should be present in the database for each post-baseline core with a prostate cancer diagnosis. For a given biopsy the % core involved will be the average value computed across all positive cores in the biopsy (i.e., average of 100*volume of cancer / total volume sampled). If there is no prostate cancer diagnosis for a given biopsy then % core involved = 0. Hence this allows summarization for both all subjects and for a "Prostate Cancer" population. Unfortunately there have been issues with these data recently (many missing values and some inconsistencies). Hence I suggest that you hold off on reporting these until the data issues are resolved. (Matt's e-mail: Tue, 16 Aug 2005 09:56:58 -0400).
7. Treatment alteration scores - these should be present in the database for all post-baseline biopsies and so can be summarized over all subjects. Note that these are provided on both a per-core and a per-biopsy (overall) basis. For analysis purposes use the overall values (OVNTASC=overall nuclear treatment alteration score (0-3), OVATASC=overall architectural treatment alteration score (0-3), and overall treatment alteration score (0-6) which is OVNTASC+OVATASC) (Matt's e-mail: Tue, 16 Aug 2005 09:56:58 -0400).