Survival

Prostate cancers are the most common cancers diagnosed in American men. Prostate cancer is also lethal. It causes more deaths annually than any other tumor except lung cancer. The vast majority of diagnosed prostate cancers do not cause death and a large majority of men diagnosed with prostate cancer ultimately die of other causes, generally related to diseases of aging including COPD and cardiovascular disease. There has been significant debate in the United States in the past several years over the need to diagnose and treat prostate cancers and the concern for overtreatment of prostate cancer. While there is some merit in this concern, the simple fact remains: prostate cancer kills some but not others. The problem is determining which is which and how to respond. Before informed treatment decisions can be made, it is important to assess the likelihood that a given tumor will recur, progress and threaten life and quality of life. The means to do this are not trivial. Life expectancy at diagnosis is a function of age, comorbidities and genetics. Assessing the likelihood of disease progression and recurrence has taken on numerous forms over the years, more recently being divided into low risk, intermediate risk and high risk (D'Amico, Data Farber), and further subdivided into very low, low, low intermediate, high intermediate and high risk (NCCN). The risk stratification is primarily based on Gleason score, PSA and T-stage. Even then it is not clear that all prostate cancers in the higher risk categories will pose a threat to life and quality of life, nor is it certain that a "low risk" prostate cancer will not pose a threat to life and quality of life.

The generic risk classification is limited in that it has traditionally only considered Gleason Score, PSA and clinical stage, while omitting other factors including PSA doubling times, extent of disease within the prostate. MSKCC and Kattan have produced nomograms which allows determine risk scores.

The CAPSURE database has been used by UCSF to develop the CAPRA score which the authors believe is easy to calculate and has comparable accuracy to the Kattan nomograms. It predicts an individual's likelihood of metastastases, cancer specific mortality and overall mortality. This index is generated by scoring points as follows:

The point score is used to delineate the low risk (score 0 - 2), intermediate risk (score 3-5) and high risk (score 6-10). The UCSF group suggests that very low risk cancers are treated well with active surveillance, low to intermediate risk are well treated with surgery or radiation alone, and intermediate to high risk require multi-modal (ie surgery plus radiation, radiation + androgen deprivation) and very high risk with radiation + androgen deprivation. Using the CAPSURE data, UCSF reports that the risk approximately doubles for every 2 point increase in score. UCSF states updates from CAPSURE demonstrate that this scoring system accurately predicts adverse events from any treatment modality.

While the CAPRA data is informative on overall risk the data presented above poses two problems. First it is a surgical series with follow on salvage therapies. It does not discuss the relative effectiveness of alternative treatments including radiation therapy. As such it is limited in its scope. MSKCC published web based nomogram calculators that several years ago compared radiation to prostatectomy, then removed the radiotherapy calculations from their web site without explanation. The second issue is that the 5 year interval is often too short an interval to identify differences in biochemical failures, salvage failures, progression of disease and death. Given the natural history of prostate cancer and the increasingly available treatments, as well as studies in other non-prostate diseases, perhaps 8-10 years is a more realistic survival timeframe, although given that, a rise from 1% to 3% at 5 years is ominous. One study indicated a substantially higher rate of prostatectomy failures at 5 years in intermediate and high risk cancers, compared with radiotherapy alone.

The Karolinksa/Gothenberg Comparative Effectiveness Review, 2014

Another study published in the British Medical Journal attempted to examine the difference in outcomes with surgery v. radiation. There is a dearth of intercomparitive data in prostate cancer. This study from Karolinska and Gothenberg asserts that biochemical recurrence is not a reasonable measure as the definitions of biochemical recurrences vary between radiotherapy and surgery and even between radiotherapy studies. The authors go on to state that the best measure of success is death from prostate cancer as an endpoint because not all post-prostatectomy patients with biochemical evidence of persistence/recurrence will progress to clinical signs and symptoms of progressive disease. Unfortunately, this study selected a cohort with significantly worse disease, greater comorbidities and older population in the radiotherapy arm over the surgical arm, and from this concluded that higher risk prostate cancer in younger men benefited more from prostatectomy than radiotherapy, taking only into consideration cause specific death. The authors conclude,

There are significant problems with these conclusions including the change in radiation treatment techniques and doses over the timeframe of the study were not accounted. Further, they did not distinguish secondary or salvage radiation therapy in their report, which would likely bias the study in favor of surgical management for all prostate cancers. In the study group 4 (Groups 1-3 were local low, intermediate and high risk prostate cancer, respectively, group 4 was metastatic), they found no difference in mortality regardless of surgical or radiation management.

The NCCN

The NCCN has recently revised its guidelines (1.2014), giving greater recommendation to radiotherapy as a primary modality treatment in high risk prostate cancer as a Category 1 recommendation (objective high level evidence supports the use of primary radiation therapy over prostatectomy except in certain very specific limited disease). For very low risk disease (GS ≤ 6, PSA < 10, < 3 cores positive and < 50% of those cores involved and PSA ρ < 0.15 ng/ml/g) either active surveillance or radiotherapy (most commonly with brachytherapy) or radical prostatectomy if life expectancy ≥ 20 years. If life expectancy is 10-20 years then active surveillance. If life expectancy is < 10 years, observation is sufficient.

For low risk disease (T1-T2a, GS ≤ 6 and PSA < 10 ng/ml and life expectancy > 10 years) then active surveillance, radiation therapy (brachytherapy or external beam treatmetn) or radical prostatectomy is reasonalbe. For life expectancy less than 10 years, observation alone is felt sufficient with no disagreement. The NCCN defines observation alone as "monitoring the course of disease with the expectation to deliver palliative therapy for the development of symptoms or a change in exam or PSA that suggests symptoms are imminent."

For intermediate risk disease, NCCN recommends prostatectomy and pelvic lymph node dissection if the predicted probability of lymphatic involvement (Kattan nomograms or Roach 2/3 PSA + 10(GS-6) ), which by definition includes nealry all intermediate risk prostate cancers, provided life expectancy is > 10 years. Alternatively, brachytherapy or external beam radiation therapy with short term androgen deprivation therapy are alos recommended. NCCN recommends brachytherapy or external beam even in cases where life expectancy is < 10 years and in cases where post-prostatectomy analysis identifies risk factors. For radiation of nodes, UCSF recommends treating once the risk rises above 15%.

For high risk and very high risk, the NCCN recommends as Category 1, based on high level evidence, radiotherapy and long term (2-3 years) androgen deprivation therapy.

For biochemical failure, androgen deprivation is the general initial therapy, either with GNRH agonist/antagonists or testosterone antagonists (bicalutamide) or a combination. More recently for symptomatic bone metastases absent extraosseous disease, ²²³Radium has been shown in the ALSYMPCA trial to be highly effective and increased median survival.

In changing its recommendations for greater use of radiotherapy as primary treatment of prostate cancer, the NCCN notes that significant increases in technology permitting highly conformal treatments and dose escalation have improved both outcomes and side effect profiles. IMRT decreases the rates of salvage therapy through dose escalation without increasing morbidity of treatment, particularly when daily image guided positioning using fiducials or CBCT is used. Kuban reported significantly better outcomes with the use of 78 Gy over 70 Gy which was even more pronounced in higher risk prostate cancer. Kuban reported a biochemical freedom from failure of 78% at 78 Gy v. 59% with 70 Gy in low risk disease and 78% v. 39% in intermediate risk disease.

The NCCN states that prostatectomy is apporpriate for anyone whose tumor is confined to the prostate. This recommendation is inconsistent with its recommendation that prostatectomy be considered in men who have a greater than 2% risk of pelvic lymph node disease in intermediate risk prostate cancer. It remains to be seen whether or not the discussion updates will the NCCN position more clear and the reasons for that position.

Concerning determination of outcomes in the post prostatectomy settings, biochemical indicators are the earliest source of potential recurrence. Three categories of biochemical recurrence/persistence are noted: a.) Failure of PSA to drop to undetectable levels; b.) PSA drops to undetectable levels but then rises on two consecutive occasions; c.) PSA fails to drop to undetectable levels but then remains low and stable for protracted times. The third category could be a result of residual prostate tissue, slow or low PSA production or residual but smoldering cancers and is infrequently seen. This third category may be safely followed until PSA rises. Salvage radiation treatment is most effective when PSA (in categories a/b) is ≤ 0.5 ng/ml and may be most effective in short PSA doubling times.

Determining the outcomes in post radiation settings, the Phoenix consensus (ASTRO) considers a rise in PSA of 2 ng/ml above post-treatment nadir as definitive of biochemical failure. Post surgical series recommend a determinant of 0.2 ng/ml as a measure of biochemical failure. Frank Critz published in the Journal of Urology (J. Urol. 2013:189 No. 3, pp 878-883) data on combined external beam and brachytherapy treatment and reported 10,15,20,25 year survival in transperineal implanted prostates of 79%. While this does not apply to RT ± ADT treatments using modern techniques, it does demonstrate the equivalence of radiotherapy and surgery when similar biochemical recurrence strategies are used to determine when and if recurrence occurs.

Intermediate Risk: RT ± ADT

Tony D'Amico studied short term androgen deprivation using neoadjuvant/concommitant and adjuvant treatment for 6 months at 70 Gy. By contemporary standards, 70 Gy is too low a dose. More commonly we use > 78 Gy today. Despite this limitation, he found that RT + ADT decreased the risk of death by half at 5 years as follows: