Primary cryoablation for Gleason 8, 9, or 10 localized prostate cancer: Biochemical and local control outcomes from the Cryo OnLine database registry

Introduction and Objective: Introduction and Objective: The increased use of cryoablation as an initial treatment for localized high-grade prostate cancer has been due to many factors including reports that cell kill from exposure to cryogenic temperatures is independent of cellular dedifferentiation and Gleason score. The objective of this study is to report the outcomes of primary cryoablation when used to treat Gleason 8, 9, or 10 localized prostate cancer at a large number of centers. Materials and Methods: Materials and Methods: Data from 1608 patients who underwent primary cryoablation at 27 centers were collected using the Cryo OnLine Database (COLD) registry. This analysis considers only the 77 patients who had a Gleason score of at least 8 and a minimum of 24 months of follow-up. Biochemical failure was de (cid:222) ned according to both the original ASTRO de (cid:222) nition (three rises) and the 2006 updated ASTRO (Phoenix) de (cid:222) nition of nadir + 2. Biopsy was performed at the physician(cid:146)s discretion, but most commonly if a patient had a rising or suspicious prostate speci (cid:222) c antigen (PSA). Results:Results: The average age at treatment was 69.6 – 8.2 years. Pretreatment PSA was 16.2 – 17.9 ng/ml and the average Gleason was 8.5 – 0.6. Patients were followed for 39.0 – 18.8 months (range: 24-120 months) and 5-year follow-up was available for 12 patients. Eight-seven percent of the patients achieved a PSA nadir < 0.4 ng/ml. Five-year actuarial biochemical survivals was 64.4 – 6.0% and 44.6 – 8.0% for the ASTRO and Phoenix de (cid:222) nitions, respectively. A total of 47 underwent posttreatment biopsy. Of these, 12 showed evidence of disease resulting in a positive biopsy rate for those who underwent biopsy of 25.5%. This yields a positive biopsy rate of the entire population of 15.6% (12/77). Conclusions:Conclusions: Cryoablation, as a primary treatment for high-grade Gleason prostate cancer practiced over a wide spectrum of users provides de (cid:222) nable biochemical and local control for a hard to manage patient population with aggressive disease.


INTRODUCTION
Despite the well-recognized stage migration that has led to an improvement in prostate cancer outcomes since the era of prostate specific antigen (PSA)based prostate cancer screening began, patients with high-grade prostate cancer remain at significant risk of morbidity and mortality. [1] Patients with Gleason score ≥8 cancer are often treated with radical prostatectomy or radiotherapy combined with adjuvant androgen blockade. Some patients have confounding factors limiting their ability to undergo surgery, including age, anesthetic risks, or simply refusal of surgical intervention. External beam underwent cryoablation as their primary therapy. Patient data were collected with the Cryo OnLine Database (COLD) registry.

MATERIALS AND METHODS
The COLD registry is a secure web-based registry consisting of case report forms designed to collect relevant pre-and post-treatment information for patients undergoing prostate cryoablation. A central institutional review board's approval covers the collection and analysis of deidentiÞ ed data, additional IRB approval for individual centers has been obtained when required by institutional policy. Investigators who have entered data into the COLD registry are listed in the acknowledgments.
Only patients with Gleason score 8 or greater having minimum 24 months follow-up were included in this analysis. Patients undergoing cryotherapy following failure of prior deÞ nitive radiotherapy or brachytherapy, and those that had undergone focal, nerve sparing, or nerve warming treatment were excluded. A total of 77 patients from 27 centers were identiÞ ed meeting the criteria.
Biochemical failure was deÞ ned according to both the original ASTRO definition of three consecutive rises following nadir (referred to as "ASTRO"), and the second ASTRO deÞ nition of nadir + 2 (referred to as the "Phoenix" definition). Biopsy was performed at the physician's discretion, but most commonly if a patient had a rising or suspicious PSA.
Kaplan-Meier analysis was used to generate curves showing the probability biochemical failure as a function of time. Statistical analysis was performed with a commercially available statistical software package (MedCalc, Mariakerke, Belgium).

RESULTS
Data were collected from 27 investigators. The average age of the 77 patients who met the inclusions criteria was 69.6 ± 8.2 years. Pretreatment PSA was 16.2 ± 17.9 ng/ml and the average Gleason was 8.5 ± 0.6. Patients were followed for 39.0 ± 18.8 months (range: 24-120 months) and 12 patients had a follow-up of at least 60 months. Posttreatment, 87% of the patients achieved a PSA nadir < 0.4 ng/ml. The 5-year actuarial biochemical survivals are 64.4 ± 6.0% and 44.6 ± 8.0% for the ASTRO and Phoenix deÞ nitions, respectively [ Figure 1]. After treatment, 47 underwent biopsy. Of these, 12 showed evidence of disease. The positive biopsy rate, for those who underwent posttreatment biopsy was 25.5% and the rate for the entire population was 15.6% (12/77). No Þ stulas were identiÞ ed in any of the patients.

DISCUSSION
The COLD registry project is designed to address the relative paucity of published data regarding cryotherapy for prostate cancer. It is notable, however, that opinions regarding the quality and quantity of data published on radical prostatectomy, external beam radiation, and active surveillance are perhaps overstated. For example, the Kaplan-Meier 5-year biochemical disease free survival data for high-risk prostate cancer managed with external beam radiotherapy as shown in Campbell's Urology textbook is based upon a surprisingly low six patients with at least 60 months follow-up. Similarly, the present series reports only 12 patients with results at 5 years. We believe that both datasets are inadequate to make solid declarations of long-term efÞ cacy. That being said, the authors do feel it is important to realize that the data supporting any form of treatment for the high-risk patient is lacking [1] and that the concept that radiation therapy is clearly established as a more efÞ cacious treatment for high-risk disease must be scrutinized.
Emerging technologies are often reserved in initial series to low-risk patients in whom failure to control disease carries less risk of metastasis and death. For example, most initial brachytherapy series are heavily weighted toward patients with low-risk disease. [4] Moreover, even recent radical prostatectomy and radiation series include more low risk than high-risk patients, but this is at least partially due to high-risk disease being less common in screened populations. [5,6] In contrast, early pioneers of cryotherapy advocated the position that if lethal cold temperatures are achieved, uniform necrosis results and cell death occurs irrespective of cellular differentiation or Gleason score. [2] In addition, publications suggesting its potential to treat locally advanced disease has led to the concept of freezing beyond the prostatic capsule to eliminate extraprostatic disease that would have resulted in a positive surgical margin had radical prostatectomy been performed. [7] As a result, cryoablation is often utilized for the treatment of high-risk localized prostate cancer. Several publications of intermediate-term results have suggested that the efÞ cacy of the procedure appears to be equivalent to radiation and surgery for lowrisk prostate cancer and potentially more efÞ cacious for moderate-and high-risk prostate cancer. [8][9][10][11] A summary of the efÞ cacies reported in the four publications reporting 5 year outcomes exclusively for primary cryoablation is presented in Table 1.
As of yet there is no universally agreed upon deÞ nition of biochemical failure following cryotherapy, but it should be noted that consensus deÞ nitions also remain relatively elusive following radiation or surgery. Radical prostatectomy completely removes the prostate and should theoretically yield an undetectable PSA. However, some residual PSA remains detectable in many patients due to a small number of remaining benign glands. The American Urological Association (AUA) Guidelines Panel recommended in 2007 that the deÞ nition of biochemical success does not require an undetectable PSA, and that failure should be a PSA > 0.2 ng/ml conÞ rmed by a second PSA reading > 0.2. [12] Further, the AUA panel found 166 different published deÞ nitions of biochemical failure, including 99 for patients undergoing radiation failure. In 1998 the Þ rst ASTRO deÞ nition of three consecutive rises, backdated to the midpoint of the Þ rst rise, was agreed upon by a consensus panel and became standard practice. It is the ASTRO deÞ nition that has been used most commonly following cryoablation. This is based on the similarity of cryoablation and radiation therapy in that both modalities result in some residual prostatic tissue and low but measurable PSA levels in most patients. The second ASTRO deÞ nition, dubbed the "Phoenix" deÞ nition, remains somewhat controversial in the eyes of many urologists.
Although it was speciÞ ed that it was not intended for use with patients undergoing cryotherapy, we have included our results using this as an additional deÞ nition solely for comparative purposes. [13] Early cryotherapy experience made it clear that speciÞ c thresholds may be meaningless due to PSA production by residual tissue surrounding the urethra or in a benign median lobe, and the fact that a PSA of 0.4 is expected when 1 g of prostate tissue has been preserved in men free of prostate cancer. [14] It is intended that the data set collected with the COLD registry will be used to create a scientiÞ cally based deÞ nition of biochemical failure that is speciÞ c to primary prostate cancer cryoablation as these data accumulate and mature.
The use of negative biopsies as a surrogate for disease control is also controversial. Regardless of the intervention, sampling error underestimates disease when using biopsy as a surrogate. This is especially likely when small volume disease is present, as evidenced by the 61% of men known to have prostate cancer that have a negative repeat sextant biopsy when evaluated on an active surveillance protocol. [15] Seventeen percent of patients with known prostate cancer on an active surveillance protocol have a negative biopsy even when saturation biopsy is performed. [16] Some authors suggest that histological evidence of malignancy identiÞ ed on biopsy should not be regarded as cancer in some postradiation settings. [17] In contrast, following cryoablation, histological results tend to fall into one of three deÞ nitive categories: Þ brous tissue (scar) indicating tumor eradication, benign prostate tissue, or prostate cancer. [18] When cryotherapy was Þ rst investigated most patients underwent posttreatment biopsy to conÞ rm local control. However, due to high-negative biopsy rates (82-98%) [8,9] most practitioners now utilize biopsy only to investigate suspicious PSA values or patterns. [19] It is not possible to know what the positive biopsy rate would be for those not determined to warrant biopsy by the treating physicians, but previous series have suggested that the likelihood of residual disease is low as demonstrated in the table.
The primary limitation of this series is the retrospective nature of a registry. There is the potential that patients with unfavorable features are not voluntarily reported to the registry, or that unfavorable outcomes are inaccurately reported. Our experience in dealing directly with the physician members and with internal audit is that we have found no evidence that there is any case selection, and to our knowledge all cases of the enrolling physicians are included. In addition, a wide variety of surgical techniques is possible with the large number of institutions participating. However, a primary goal of the COLD registry is to determine outcomes in "the real world" without the inherent reporting bias of single surgeon series. These data suggest that bDFS and complication rates are consistent with earlier single-center reports.