The present study is the first CaP survival analysis using the Ohio CALD, which mirrors in structure the SEER-Medicare files. We evaluated CaP incidence data linked with Medicare and Death Certificate files, and concluded that the definitive treatment modalities, RP and BT, were associated with significantly better overall and disease specific survival compared to NT. The present results provided demonstrate that the risk of CaP specific deaths decreased more than 50% when receiving BT alone or in conjunction with EBRT or ADT compared to NT, after controlling for age, race, comorbidities, tumor stage, and Gleason score. There are limited data available through randomized clinical trials providing information that directly compares all the treatment modalities: RP, BT, EBRT, and ADT versus NT (3
). The overall survival rates and the definitive treatment effects of the present study are in agreement with the previously published studies using SEER-Medicare data analyzing aggressive treatment versus non-aggressive treatment for localized disease (13
), and distant disease (14
) – findings that attest to the validity of the data represented in the Ohio CALD.
Studies investigating the effect of RP versus NT in predicting overall and disease specific survival have not yielded consistent results. Findings from a randomized clinical trial (n=695 men with localized disease) indicated that RP predicted a significant overall survival benefit, in addition to a very significant reduction in the risks of disease specific death, metastasis, and local tumor progression comparing to watchful waiting (20
). There were two other randomized clinical trials comparing RP versus NT or RP versus EBRT. With small sample size, neither trial was able to convincingly demonstrate an advantage of RP over NT or RP over EBRT (21
). A retrospective population-based cohort study using data from the Connecticut Tumor Registry (n=767 men) showed that the annual mortality rate from CaP remained stable 15 years post diagnosis, a finding that does not support aggressive treatment for localized low-grade CaP (22
). The present study provides evidence supporting RP for localized cohort in a more heterogeneous population.
In this study, we compared BT versus NT and showed that BT was significantly associated with better overall and disease specific survival. The risk of CaP specific death decreased more than 50% when receiving BT alone or in conjunction with EBRT or ADT compared to NT. No randomized clinical trials or large cohort studies comparing BT versus NT were found when performing Medline search. BT and NT are usually being offered to men with localized disease especially with low Gleason score. However, as we mentioned before, these men may carry undetected high-grade disease (23
). Men with shorter life expectancy tended to be observed instead of being treated. However, even after adjusting for age, race, comorbidities, tumor stage, and Gleason score, we still conclude that BT yields significantly better survival outcomes than NT. Baseline PSA and its velocity have a critical role in determining treatment, unfortunately PSA values cannot be obtained from the currently available databases. However, this is essential when comparing NT with other definitive treatment regimens because NT is usually assigned to those with low PSA, low Gleason score, and surrogates for clinically insignificant CaP (24
). We would expect even more significant effects comparing both RP and BT versus NT when controlling for pretreatment PSA. This is especially true when comparing EBRT versus NT. Recent hospital-based or multi-center studies showed that 3D conformal radiotherapy and intensity-modulated radiation therapy (IMRT) could provide better disease specific survival, compared to the conventional radiation therapy; and that the outcome of EBRT and BT was similar to those of surgery patients with low risk disease (26
). The present study showed significant decreased risk for patients treated with EBRT either alone or in conjunction with ADT compared with NT for the entire cohort. A more than 40% decreased CaP specific risk for the entire cohort indicated using EBRT could significantly improve survival in a large heterogeneous population. It failed to conclude whether EBRT with or without ADT, could provide significant benefit for disease specific survival for localized disease without adjusting for pretreatment PSA. One reason may be that the follow-up period is not long enough. With a small number of disease-specific deaths in localized disease, it is more difficult to show significant differences in survival. Another reason is that men who received EBRT usually have more aggressive disease than those who do not initiate definitive treatment within 6 months. In addition, even though we were able to control for the most important tumor and patients’ characteristics, the results may be confounded by variables that we could not account for in our analyses. These include preoperative PSA and patients’ performance status. For localized disease, it would be expected that those who had NT or received BT or RP as mono-therapy, would have a lower PSA than those receiving EBRT. EBRT in conjunction with ADT or ADT alone provides better survival than NT in distant disease group, which is in agreement with the previous SEER-Medicare study by Lu-Yao, et al (14
The majority of patients who show progression are often patients in whom occult metastatic disease may have been present prior to definitive therapy, and thus they were under-staged, and technically not curable. Recent introduction of advanced imaging techniques (27
) may hold promise for selecting patients for local therapies, or help to determine which patients would benefit from the addition of systemic treatment with either adjuvant hormonal therapies or other forms of adjuvant chemotherapy, along with improved local therapies like cryotherapy, cyberknife, and other new EBRT techniques such as IMRT and 3D conformal radiotherapy, etc. might be helpful in increasing the disease free survival of CaP patients (28
Despite our attempts to adjust for all available confounding factors in the regression model, without proper randomization, the treatment effect may be biased with known and unknown confounding factors. Pretreatment PSA >10 or >20 ng/mL, is one of the most important factors when deciding the treatment modality, though not specific for CaP, it is usually considered a risk factor associated with disease specific survival (2
). Recent studies show that PSA velocity not only predicts the presence of CaP but also is highly associated with disease specific death (29
). Percent positive biopsies and proportion of higher Gleason grades are some other factors (30
). Many other factors like smoking, certain food consumption, and other life style may significantly impact treatment selection and survival, which could not be adjusted for (32
Additionally, although we controlled for comorbid conditions, there could well be other conditions such as performance status, functional dependence and presence of geriatric conditions, which may characterize a patient as unfit to undergo more aggressive treatment. Future studies that link large population based data with multi-center clinical data are a potential direction when more, higher quality electronic clinical data may be routinely available. Like SEER data, the OCISS updates the stage information after a surgery. For CaP, RP is the only procedure that may have an up-graded stage. Compared to other treatment methods, RP may potentially have higher survival controlling for stage. However, we expect this proportion to be low and some even report the down grade of Gleason score (35
). In the case that this happens and there is some degree of differential bias, we could expect better survival after BT than RP after controlling for this effect. For the finding that RP is better than NT, the hazard ratio will migrate towards the null when considering the up-grade effect.
As expected, age, tumor stage, Gleason score, and patient comorbidities are all very important factors in predicting overall and disease specific survival. We conclude that in the CaP survival analysis, without proper adjustment for age and the patient’s health condition, the treatment effect cannot be evaluated conclusively.
In summary, the findings of this study suggest RP and BT significantly improve both overall and disease specific survival compared with NT, especially when used in conjunction with EBRT and ADT. This study provides more evidence supporting that earlier initiation of definitive treatments like BT or RP is better than NT in a more heterogeneous population. Despite our current inability to determine pre-treatment PSA data from our dataset, it should provide a sense of caution in not recommending treatment to patients who appear to have some higher degree of risk from the clinical staging, including not only PSA level, but PSA doubling times, Gleason scores of 7 or greater, and perhaps disease volume based on % positive core biopsies and % positive of each core. These results are very important for clinicians, patients, and their families in deciding the next step when facing the new diagnosis. Further research on CaP treatment related urinary and bowel function and erectile dysfunction are needed to better evaluate and interpret treatment outcomes.