Impact of the percentage of positive prostate cores on prostate cancer-specific mortality for patients with low or favorable intermediate-risk disease
Authors: D'Amico AV, Renshaw AA, Cote K, Hurwitz M, Beard C, Loffredo M, Chen MH
Source: J Clin Oncol. 2004 Sep 15;22(18):3726-32.
Risk stratification using pretreatment markers is used to help determine treatment strategies in prostate cancer . Patients are generally divided in low, intermediate and high risk of recurrence based on stage (T stage for locally confined disease), pathology (Gleason Score), and serum markers (PSA levels). PSA has been an invaluable tool in not only screening for prostate cancer, but also in risk stratification and follow-up of patients. PSA elevations have been correlated with larger or more aggressive cancers, and thus PSA is a good pretreatment risk stratification marker. Due to the power of PSA in predicting relapse, many studies have used PSA failure (or increase after treatment) as a surrogate marker of cancer recurrence. In is generally agreed upon that patients with Gleason Score (GS) of 8 or PSA > 20 are considered to have a high risk of relapse. Similarly, Gleason Score of 6 or less and PSA < 10 are considered markers of low risk carcinomas. The question however becomes how to stratify intermediate risk patients and how these pretreatment markers may suggest alternative treatment strategies for these intermediate risk patients. Attempted stratification strategies of these patients have looked at a number of factors including: Gleason Score of 7 being 3+4 vs 4+3; PSA >10 and up to 15ng/ml vs greater than 15 ng/mL; percentage of positive cores on biopsy; greatest percentage of one core involved by carcinoma; and the total linear millimeters of carcinoma. In an effort to answer this question, and using a more concrete end point than PSA failure, D'Amico et al. have asked a simple question: For low to favorable intermediate-risk patients, how does the percentage of positive cores on biopsy correlate with disease specific mortality?
- Prior study D'Amico et al . (JCO 21:2163-2172, 2003) looked at 7,316 status post prostatectomy or radiation and looked at pretreatment risk groupings and supported the predictive nature of pretreatment markers including age, PSA, GS and T stage. This study looked at the predictive nature and disease specific mortality in the PSA era.
- Two prior D'Amico studies looking at % positive cores as predictors of outcome [as measured by PSA failure] after prostatectomy or radiation (JCO 18:1164-1172, 2000 and IJROBP 49:679-684, 2001). These two studies showed clinical utility of using a breakpoint of 50% for percentage of positive cores (% PC) and thus this breakpoint was used in this study. This current study was designed to determine if that endpoint of PSA failure is meaningful when looking at prostate specific mortality after XRT monotherapy.
- A prior study showed a direct association between % PC and tumor volume (Stamey et al . JAMA 281:1395-1400, 1999)
D'Amico et al . tested pretreatment factors as predictors of outcome after XRT (external beam radiation therapy) in this retrospective study.
Material and Methods
- 421 patients treated between 1988 and 2002 at a single institution (St. Anne's hospital in Fall River , MA ) with no patients lost to followup.
- Median age 71.9 yrs (range 44.9 89.6 yrs)
- Low/favorable intermediate-risk pts
- Low risk: T1c-T2, PSA level d 10 ng/mL and biopsy Gleason Score d 6
- Favorable intermediate-risk: T1c-T2 with PSA >10 to 15ng/ml OR Gleason Score 3+4
- NOTE: Included pts with T2c disease in low risk groupings (AJCC 2002 staging)
- Exclusion: GS 4+3, PSA > 15ng/mL, evidence of perineural invasion, prior TURP (transurethral resection of the prostate).
- All path centrally reviewed by a single pathologist
- Minimum and median number of cores sampled was 6 with range of 6 to 16 cores.
- Radiation Treatments
- Low risk: median dose 70.4 Gy after 95% normalization in 180 cGy fractions to prostate with 1.5cm margin.
- Favorable intermediate-risk: 45 Gy in 180 cGy to prostate plus seminal vesicles with 1.5 cm margin with cone down to low risk fields and dosing thereafter.
- ALL doses reported as minimum dose of PTV (planned target volume)
- Median f/u 4.5 yrs (range 0.4 to 13 yrs) with 117 (28%) of patients dying.
- Cause of death: For prostate specific mortality pt required to have progressive hormone-refractory metastasis prostate cancer. To be scored required positive bone scan, progression despite ?exhausting all known hormonal manipulations' and having at least seen chemotherapy. Note: This is a very narrow definition.
- Cox regression multivariable analysis performed (PSA level, Gleason score, T stage, or % PC [percentage of positive cores]) to predict time to prostate cancer-specific mortality using a two-sided log-rank P value. Compared:
- PSA d 10 vs 10-15 ng/mL
- GS d 6 vs GS 7 (3+4)
- T1c and T2a vs T2b and T2c
- % PC less than 50% vs e 50%
- 117/421 (28%) pts died with only 15/117 (13%) from prostate cancer.
- % PC was the ONLY significant predictor (P d 0.03) of Disease Specific Survival/Mortality
- At 8 yrs (Note: median f/u only 4.5 yrs) 5 to 9% of those with e 50% PC had disease specific mortality vs 1% if less than 50% PC.
This paper is an important step towards determine treatment strategies for these low to favorable intermediate-risk patients. A majority of patients diagnosed in the PSA era would qualify for this study, thus increasing its relevance. One criticism of many studies on the topic of prostate cancer has been the use of a surrogate endpoint of PSA failure. Critics have argued that PSA failure is simply a blood marker and disease specific survival and overall survival are far more important outcomes. In answer to that criticism, this paper looks at disease specific mortality as its endpoint. This strength however is also its weakness in that only 15 patients qualify for analysis having died with progressive hormone-refractory metastatic prostate cancer. D'Amico et al . have used a rather narrow definition and have failed to comment on overall survival in this study. Interesting points raised include the possibility that with more follow-up a PSA level above 10, but below 15 may also approaches significance (p=0.11). The reasons for worse outcome with greater %PC is hypothesized as being due to inadequate local control with larger tumor volumes and the possibility of larger tumors having unrecognized micrometastases. With a greater number of cores positive there is a higher likelihood of occult seminal vesicle involvement or even under sampling causing erroneously low Gleason Scoring. This paper is the first to show that percentage cores positive is an important consideration with respect to mortality in these early stage prostate cancers. Further studies assessing the role for dose escalation or hormonal ablation in these patients will hopefully pave the way for improved treatment strategies.