During the retropubic era, the common definition of treatment failure for prostate cancer was clinical recurrence based on the digital rectal exam or the development of radiographic evidence of metastasis. In 1997, the American Society of Therapeutic Radiology and Oncology (ASTRO) convened a consensus panel that developed the current standard definition of biochemical failure 69. While this definition was specifically developed for patients undergoing EBRT, it has been employed in many series reporting outcomes following brachytherapy. The consistent use of this definition allows for direct comparisons between EBRT and brachytherapy. Biochemical failure was defined by the ASTRO consensus panel as three consecutive rises in the PSA value, obtained at 6-month intervals with a minimum follow-up of 24 months 69. The time to failure is defined as the mid-point between the PSA nadir and the first PSA rise. Unfortunately, until the ASTRO consensus definition was developed, the literature reporting outcomes from modern permanent prostate brachytherapy used various biochemical definitions making comparative interpretations of the data difficult.
Five-year biochemical freedom from recurrence following modern permanent brachytherapy is reported between 75-100% with 8-13 year results between 66-88% 70, 71. In the largest multi-institutional series of over 3500 patients, Kattan et al. report a five-year 79.8% biochemical freedom from recurrence rate for those patients treated with either monotherapy or EBRT and implant (Figure 4) 72. No patients treated with AD were included in that study. Multivariate analysis from that series indicated that the pretreatment PSA value and the Gleason sum were highly significant factors predicting biochemical success (p=0.0001 and 0.0003, respectively), while the addition of EBRT was significant with a p value of 0.0487.
Univariate and multivariate analysis from most studies on patients with clinically localized prostate cancer treated with either permanent brachytherapy or EBRT concur that the pretreatment PSA values and Gleason sum are highly significant factors predicting biochemical success 73-75. As a result, several methods for developing risk stratification schemes have been proposed, each of which appear to distinguish significance relative to biochemical freedom form survival (Figure 5) 72, 76.
While the concept of risk stratification is intuitive and simple to comprehend, recently developed nomogram analysis allows one to predict outcome better 77. While somewhat less intuitive, the nomogram approach assesses variables such as the pretreatment PSA value and Gleason sum in a continuous manner and independent of other factors to predict biochemical disease control. (Figure 6) 78. Use of additional pathology data such as percent positive core biopsy specimens has been identified to further enhance the ability to predict outcome. Potters et. al. examined information from brachytherapy patients identifying 26 independent variables 34. This additional information predicted biochemical freedom from recurrence better than that of the nomogram model. As such, the concept of principle component analysis used in the Potters study to evaluate and compile multiple pathology variables will be used to construct a second-generation nomogram for patients considering brachytherapy.
Implant Dosimetry and Outcomes:
The first dose response paper for modern permanent prostate brachytherapy reported from Stock et. al. analyzed the results of CT-based post-implant dosimetry (using TG-43 guidelines) in 134 patients treated with I-125 implants for T1 to T2 prostate cancer 13. Increasing D90 values from <100Gy, 100-119.9 Gy, 120-139.9 Gy, 140-159.9 Gy, and >160 Gy were associated with improved freedom from PSA failure rates of 53, 82, 80, 95 and 89%, respectively (p=0.02) at 4 years. A dose cutoff point was found at 140 Gy, with PSA control rates of 68% for those patients receiving a D90 <140 Gy compared to 92% for those with a D90 >140 Gy (p=0.02). Likewise, Potters et al examined implant dosimetry in 719 patients 14. Using the D90 dose relative to the prescribed dose (D90%), a significant breakpoint was identified at a D90 dose at or above 90% of the prescribed dose. In that study, this breakpoint was significant for I-125 (p=0.01), Pd-103 (p=0.04), monotherapy brachytherapy (p=0.001) and the addition of androgen ablation to the implant (p=0.001). The only factor that did not demonstrate a dose effect was the addition of EBRT to brachytherapy.
Another component of implant dosimetry is the need to help predict for toxicity. Wallner et. al. 42 analyzed 45 patients treated with I-125 implantation who had CT based dosimetry performed 2-4 hours after implantation and related these findings to urinary and rectal morbidity. In patients who developed RTOG grade 0-1 urinary morbidity, an average of 10 mm of urethra was irradiated to doses >400 Gy (pre-TG43), compared to 20 mm for patients experiencing grade 2-3 morbidity (p=0.07). They concluded that both the dose and length of urethra irradiated were related to urinary morbidity. Similarly, when examining rectal morbidity, in patients developing RTOG grade 1-2 rectal morbidity, an average of 17 mm2 of rectal wall was irradiated to doses >100 Gy, compared to 11 mm2 for patients experiencing no rectal morbidity 42.