Focal Therapy in Prostate Cancer: Determinants of Success and Failure

Introduction

T he incidence of low to intermediate risk prostate cancer is rising as a result of informal and formal prostate-specific antigen (PSA) screening practices. ¹ The treatment burden from radical therapies is high with significant genitourinary and rectal toxicity. Active surveillance, on the other hand, carries surveillance and psychological burden with risk of progression. Focal therapy aims to find a middle ground between surveillance and radical therapies by treating the cancer alone, with a margin, and preserve as much tissue as is practical. ^{2 –4}

Early feasibility studies have demonstrated the absence of rectal toxicity and preservation of genitourinary function in up to 90% to 95% of men. ⁵ To date, focal therapy case series have evaluated hemiablation of unilateral disease (Gleason ≤7, PSA ≤15, ≤T2bNoMo) using cryosurgery and one using high-intensity focused ultrasound (HIFU). These have demonstrated impotence rates of approximately 15% with little to no incontinence. These have used a variety of methods to identify unilateral disease from Doppler transrectal ultrasound (TRUS) biopsies, TRUS alone, and template biopsies and have generally shown poor reporting standards because of their retrospective nature with short follow-up and generally small numbers of patients. Prospective phase II trials are under way or completed at a number of centers (Table 1). The first of these trials conducted at our center in London is completed and has demonstrated 95% preservation of genitourinary function and 90% negative biopsy rates at 6 months after focal therapy using HIFU (under submission).

A strategy to evaluate focal therapy should be embedded in trials that ascertain the key medium-term outcome measures that will determine the success or otherwise of this proposed alternative to the standard of care. This is far from easy. The ideal outcomes, metastases and death, require trials of at least 10 years in duration because of the long lead time bias inherent in the screen detected population of prostate cancer, so surrogate markers of failure have been proposed within the standards of care. Radical surgery uses a PSA threshold < 0.2 ng/mL, whereas radiotherapy uses the American Society for Therapeutic Radiology and Oncology (ASTRO) Phoenix consensus criteria of two consecutive rises above PSA nadir. ⁶ On the other hand, active surveillance uses clinical, histological, and biochemical measures of progression, with the latter far from being validated. ⁷

If focal therapy is to be proposed as a challenge to these existing strategies, it is likely that many of these measures may not be suitable. While treating the cancer foci it does so by leaving substantial amounts of prostate tissue untreated. It can neither achieve unrecordable PSA levels, nor is it easy in such a setting to apply the ASTRO-Phoenix criteria. Equally, progression as defined by active surveillance regimens may not readily translate to a man who has had all clinically significant cancer treated using hemiablation but still has 50% of the prostate still present, for instance. In addition, the optimal ablative modality to deliver focal therapy is far from clear, with cryosurgery, HIFU, photodynamic therapy, photothermal therapy, and brachytherapy all emerging as likely candidates. ⁸ Whether there will be ablative-specific tissue responses that require adjustments in outcome measures is unclear, but will need to be investigated.

Biochemical Outcomes

Conventional therapies have used surrogate endpoints, primarily biochemical (PSA) outcomes, to determine the success or otherwise of radical surgery or radiotherapy. ^9,10 Radiotherapy definitions tend to overestimate biochemical-free survival generally with a 5-year lag in deeming a treatment failed compared with surgery. Even within these long established therapies, there can be wide variation in definitions used for failure with over 166 definitions reported in the literature. ¹¹ The definition of success for ablative technologies delivered in a whole-gland manner has not reached consensus. ^12,13 Some have adopted ASTRO criteria or modified the acceptable increase above nadir to 1.2 ng/mL (deemed the Stuttgart definition). ¹⁴ Others, on the other hand, have determined that a PSA nadir upper threshold should be used, although there is no agreement on whether this should be < 0.2, < 0.4, or < 0.5 ng/mL. ⁸

With untreated tissue remaining after focal therapy, it would seem unwise to use absolute biochemical parameters or even PSA kinetics alone without a form of standardization against a particular patient's volume of untreated tissue and volume of cancer ablated. PSA will vary according to the ablative technology, whether a particular device was used, amount of tissue ablated, and any residual tissue tat remains. With the latter, the untreated tissue may be benign in its entirety or have clinically insignificant areas of low-volume low-grade cancer that have been deliberately untreated to deliver tissue preservation. Therefore, the parameters that may prove to be of greater use are discussed below.

Biopsy Outcomes

Biopsies should be used to determine absence of disease within treated areas to verify short-term focal ablative success as well as untreated areas to detect recurrent and de novo disease, respectively, in the medium to long term. However, TRUS-guided biopsies, if used for the latter objective, will be subject to the same systematic and random errors inherent in this test when applied in diagnosis of prostate cancer. Thus, a degree of targeting using noninvasive imaging to identify clinically significant lesions may be necessary. TRUS-guided biopsies used in this setting are also prone to detect clinically insignificant cancers (low volume, low grade) that are unlikely to influence disease progression—such foci may indeed have been missed on initial localization strategies. ¹⁸ Therefore, their subsequent detection many years after focal therapy need not necessarily equate to the verdict of progressive, recurrent, or de novo cancer. Definitions related to clinical significance would need to take account of grade and cancer core length involvement as in diagnostic strategies—2 to 3 mm of cancer in any one core with absence of Gleason pattern 4 may be a starting point, ¹⁹ but such criteria will need careful validation. In addition, more accurate volume assessments of cancer foci, if present on surveillance imaging, will be needed. If not observed on surveillance imaging, a posttreatment template transperineal-mapping biopsy may be required to determine the disease burden of any cancer found on surveillance biopsies. ²⁰

Biopsies will need to also take into account the therapeutic strategy employed at baseline. Was the strategy one of ablating all measurable diseases with absence of any cancer in untreated areas or was some element of cancer accepted in untreated areas, that is, absence of any clinically significant areas (up to 2–3 mm low-volume low-grade foci accepted)? In addition, were template-mapping biopsies or TRUS-guided biopsies used to localize disease? If the latter, then disease found on surveillance biopsies may simply represent the sampling error of the initial localization tool employed.

In summary, biopsies of the prostate for surveillance after focal therapy must be used in a more refined and accurate manner taking into account the localization and therapeutic strategy used. Image-based targeting and, where necessary, template or saturation biopsies should be employed in a similar manner to that before focal therapy. The key determination will be whether recurrence or de novo cancer found on surveillance after focal therapy is clinically significant or insignificant with the latter triggering continued surveillance and the former warranting further treatment. Indeed, the need and chronology of further treatment may be an ideal parameter against which to assess focal therapy if the comparator arm is active surveillance although such a determinant of failure is less obvious when radical therapies are considered. An outcome that may have application across all therapies and active surveillance is time to hormonal therapy, and this may serve as an important outcome in comparative trials in future.

Imaging Outcomes

Imaging may have a role in the pretreatment selection of focal therapy candidates. ²¹ Early gadolinium contrast-enhanced MRI, within 1 to 2 weeks of treatment, has been shown to accurately predict areas of necrosis after whole-gland and focal HIFU as well as other modalities such as photodynamic therapy, thus having a role in early verification of treatment effect. ²² In addition, a number of authors have stated that multifunctional MRI (T2-weighted, dynamic contrast enhancement, diffusion, and spectroscopy) seems to meet the ideal attributes for detection of clinically significant cancer, thus potentially being used to drive the delivery of focal therapy. A number of groups have demonstrated accuracy of over 85% to 90% for lesions that measure 0.2 or 0.5 cc in volume. The exclusion of significant lesions may be more important in focal therapy, and hence the negative predictive value, and in this negative predictive value for 0.5 cc lesions has been demonstrated to be as high as 95% if multifunctional MRI is used before TRUS-guided biopsy. ^23,24 As 0.5 cc is commonly used as the threshold at which prostate cancer lesions become clinically significant, the inherent ability of multifunctional MRI to be able to detect large lesions and not detect small lesions may be its greatest attribute, and may serve to justify its use not only in disease localization for focal therapy but also as a triage test before TRUS-guided biopsy. ²⁵ These results, which used radical prostatectomy reference standard validation, need verification in other centers as part of multicenter trials. In addition, it would be key to also use validation against template-mapping biopsies, a reference standard that would have less inherent selection bias because of its applicability to all men.

It therefore seems logical that, in medium to long term, surveillance of untreated areas of the prostate could be undertaken by multifunctional MRI to detect recurrence of clinically significant cancer. A negative MRI would imply absence of clinically significant disease that requires no treatment. This would avoid any potential over-treatment after focal therapy. Other ultrasound imaging modalities (elastography; ultrasound tissue characterization, e.g., Prostate HistoScanning^® (AMD SAS, Waterloo, Belgium); and contrast-enhanced ultrasound) that are starting to demonstrate promise in the detection of prostate cancer before treatment could also be applied after focal therapy. ^{26 –28}

Future Perspective

Formulating a research strategy for an evaluation of focal therapy taking into account all the data that exist to date and all the uncertainties that are present in any new innovation requires discussion within consensus panels that have free access to existing treatment outcomes from the various groups that have already conducted focal therapy. We have outlined the likely pathway to evaluating this new strategy and health technology in Tables 2 and 3. These recommendations are far from prescriptive or correct but form the basis on which future discussions and debate may occur.

Conclusion

Focal therapy has reached prominence in prostate cancer research. At its heart, the discussions around a research strategy to evaluate focal therapy, before its informal diffusion, is the controversy over what is and what is not clinically significant cancer. Taking this paradigm shift forward will require an equally large paradigm shift in how we diagnose prostate cancer. As current systematic but random TRUS biopsies are inherently inaccurate for ruling-in or ruling-out clinically significant cancer, it seems prudent that imaging tools at diagnosis and in surveillance protocols will drive the success or otherwise of prostate-preserving therapy.