ÌÇÐÄvlog

Key Points

QuestionÌý What are the long-term outcomes for patients with prostate cancer whose cases are managed with protocol-directed active surveillance?

FindingsÌý In this multicenter cohort study that included 2155 individuals with a median follow-up time of 7.2 years, the 10-year incidence of upgrading at biopsy and definitive treatment were 43% and 49%, respectively. The 10-year incidence of metastasis or prostate cancer mortality were 1.4% and 0.1%, respectively. There was no significant difference in adverse outcomes in men treated within the first 2 years of surveillance vs later on.

MeaningÌý Protocol-directed active surveillance is a safe management strategy for avoiding overtreatment and preventing undertreatment.

ImportanceÌý Outcomes from protocol-directed active surveillance for favorable-risk prostate cancers are needed to support decision-making.

ObjectiveÌý To characterize the long-term oncological outcomes of patients receiving active surveillance in a multicenter, protocol-directed cohort.

Design, Setting, and ParticipantsÌý The Canary Prostate Active Surveillance Study (PASS) is a prospective cohort study initiated in 2008. A cohort of 2155 men with favorable-risk prostate cancer and no prior treatment were enrolled at 10 North American centers through August 2022.

ExposureÌý Active surveillance for prostate cancer.

Main Outcomes and MeasuresÌý Cumulative incidence of biopsy grade reclassification, treatment, metastasis, prostate cancer mortality, overall mortality, and recurrence after treatment in patients treated after the first or subsequent surveillance biopsies.

ResultsÌý Among 2155 patients with localized prostate cancer, the median follow-up was 7.2 years, median age was 63 years, 83% were White, 7% were Black, 90% were diagnosed with grade group 1 cancer, and median prostate-specific antigen (PSA) was 5.2 ng/mL. Ten years after diagnosis, the incidence of biopsy grade reclassification and treatment were 43% (95% CI, 40%-45%) and 49% (95% CI, 47%-52%), respectively. There were 425 and 396 patients treated after confirmatory or subsequent surveillance biopsies (median of 1.5 and 4.6 years after diagnosis, respectively) and the 5-year rates of recurrence were 11% (95% CI, 7%-15%) and 8% (95% CI, 5%-11%), respectively. Progression to metastatic cancer occurred in 21 participants and there were 3 prostate cancer–related deaths. The estimated rates of metastasis or prostate cancer–specific mortality at 10 years after diagnosis were 1.4% (95% CI, 0.7%-2%) and 0.1% (95% CI, 0%-0.4%), respectively; overall mortality in the same time period was 5.1% (95% CI, 3.8%-6.4%).

Conclusions and RelevanceÌý In this study, 10 years after diagnosis, 49% of men remained free of progression or treatment, less than 2% developed metastatic disease, and less than 1% died of their disease. Later progression and treatment during surveillance were not associated with worse outcomes. These results demonstrate active surveillance as an effective management strategy for patients diagnosed with favorable-risk prostate cancer.

Active surveillance emerged as a management approach for favorable-risk prostate cancer in response to overtreatment of indolent disease that resulted with the introduction of prostate-specific antigen (PSA) screening. This strategy involves monitoring with clinical examinations, serum PSA, periodic biopsies, and, more recently, multiparametric magnetic resonance imaging (MRI), such that definitive treatment with surgery or radiation is either delayed until there are signs of cancer progression or avoided altogether if no progression is detected, mitigating common adverse effects associated with treatment, such as urinary incontinence, urinary irritative symptoms, and erectile or bowel dysfunction.¹ Currently, active surveillance is the preferred management for low-grade prostate cancer.^2-4

Although several studies have demonstrated that the majority of patients undergoing surveillance have favorable long-term outcomes,^5-9 only about 60% of eligible patients in the US use surveillance,¹⁰ with uptake varying by clinician.¹¹ Reasons for this are multifaceted, but include fear of undertreatment, where a window of cure is missed in cancers that initially appear indolent but exhibit aggressive features during the course of surveillance.¹² Furthermore, current clinical guidelines provide minimal guidance for optimal surveillance regimens. To improve utilization of active surveillance, a better understanding of how to best balance avoiding overtreatment while preventing undertreatment is needed.

This study reports the long-term oncological outcomes in the multicenter, protocol-directed, and prospectively accrued Canary Prostate Active Surveillance Study (PASS) cohort. To assess active surveillance as a long-term prostate cancer management strategy, the study examined rates of disease reclassification and treatment, recurrence after treatment, and the occurrence of metastases and death from prostate cancer or other causes. The study also evaluated oncologic outcomes in patients treated after the first prostate biopsy following diagnosis compared with patients treated after several years of surveillance.

Canary PASS is a multicenter, prospective, observational study that enrolled active surveillance patients at 10 sites from 2008 to 2022 (Beth Israel Deaconess Medical Center, Eastern Virginia Medical Center, Emory University, Stanford University, University of British Columbia, University of California San Francisco, University of Michigan, University of Texas Health Sciences Center at San Antonio, University of Washington, and Veterans Affairs Puget Sound Health Care System). Interested and eligible patients with clinically localized prostate cancer whose cases were managed with active surveillance rather than primary therapy provided informed consent under institutional review board supervision (). In order to sample representative patients using active surveillance to manage their cancer at the study sites, there were no exclusions based on grade group (GG) or PSA. Clinical and pathologic data were abstracted from medical records from time of diagnosis. Participants self-identified their race and ethnicity based on fixed categories per guidelines established by the National Cancer Institute. Marital status and education were self-reported starting in 2015.

Patients were followed up using a standardized protocol in which PSA was measured every 3 months before 2020 and every 6 months starting in 2020. Prostate biopsies were protocol directed at 6 to 12 months after diagnosis, 2 years after diagnosis, and then every 2 years; biopsies within 12 months of the target date were considered compliant. Sequential biopsies are described as diagnostic, confirmatory (first surveillance), and subsequent surveillance. The cumulative incidence of having the confirmatory biopsy within 2 years of diagnosis was 88%, and 97% of participants had the confirmatory biopsy within 5 years (eTable 1 in Supplement 1). After the confirmatory biopsy, the cumulative incidence of having a biopsy within 3 years was 83% and within 5 years was 90%. Following the second surveillance biopsy, an estimated 74% and 83% of participants had a subsequent biopsy within 3 and 5 years, respectively. Other studies, including MRI and biomarker tests, that came into clinical use since the inception of PASS, were performed at the clinicians’ discretion and these data were collected. Biopsy results were read by genitourinary pathologists at study sites and offsite diagnostic biopsy results were formally reread at study sites. In a previous analysis, the study found no substantial differences in reclassification rates in those diagnosed onsite vs offsite.¹³ Clinic visits occurred every 6 months until definitive treatment, after which, clinical data (eg, tests to monitor prostate cancer, treatments, and outcomes) were collected annually. Ten percent of participants were lost to follow-up.

Outcomes evaluated for this analysis include biopsy grade reclassification, definitive treatment (radical prostatectomy, radiation, androgen deprivation therapy [ADT], other), adverse pathology at prostatectomy, recurrence after treatment, metastasis, and both overall and disease-specific mortality. Biopsy grade reclassification was defined as any increase in GG on surveillance biopsy. Extreme reclassification was defined as an increase to GG3 or above. Adverse pathology at prostatectomy was defined as GG3 and above or cancer that has spread to the seminal vesicles or lymph nodes, as denoted by pT3b and above or N1. Recurrence after prostatectomy was defined as 2 PSA values 0.2 ng/mL or above, secondary treatment with rising PSA, or cases in which the PSA never dropped to undetectable levels after prostatectomy. Recurrence after radiation therapy was defined as PSA greater than nadir +2 ng/mL or secondary treatment more than 6 months after radiation with evidence of rising PSA or positive prostate biopsy. Metastasis using the American Joint Committee on Cancer definition of disease in nonregional lymph nodes, soft tissue, or bones and cause of death were determined by independent review of the PASS end points committee.

Data collected before August 2022 were used in this analysis. Survival analyses for time to reclassification, treatment, metastasis, and mortality started at the clinically relevant time of diagnosis; analysis of recurrence after treatment started at time of treatment. For all outcomes evaluated, participants without the event of interest were censored at their last contact. For time to reclassification or time from one biopsy until the next, treatment prior to reclassification was considered a competing risk event. Cumulative incidence rates were calculated for the total cohort (N = 2155) and for a subcohort enrolled shortly after diagnosis, before confirmatory biopsy (n = 1403), using the Kaplan-Meier method to account for censoring and the Fine-Gray subdistributional assumption in the presence of competing risks.^14,15 Analyses were conducted using R version 4.1 (R Foundation).

A total of 2155 patients with localized prostate cancer and a median (IQR) follow-up time after cancer diagnosis of 7.2 (4.2-11) years were included in the study. Nearly two-thirds (n = 1403) enrolled shortly after diagnosis, prior to the confirmatory biopsy. At diagnosis, the median age was 63 years, the majority of patients (1779 [83%]) were non-Hispanic White, and 159 (7%) were Black (Table 1). Median PSA was 5.2 ng/mL, and the majority of patients were diagnosed with GG1 cancer (90%) and met traditional definitions of very low– or low-risk prostate cancer² (83%; Table 1).

Figure 1 depicts biopsy reclassification and treatment during the active surveillance continuum. There were 2008 patients with at least 1 surveillance biopsy after the initial diagnostic biopsy. About half (n = 1003) did not have reclassification or treatment. Reclassification to a higher-grade cancer occurred early during surveillance (at confirmatory biopsy) in 374 patients and later during surveillance (at any subsequent surveillance biopsy) in 404 patients. After biopsy reclassification at either confirmatory or subsequent biopsy, most patients received definitive treatment, but about one-quarter remained receiving active surveillance. Concomitantly, about one-quarter of treated patients had no associated upgrading.

To evaluate the rate of grade reclassification at confirmatory biopsy, this study examined the subcohort of 1403 participants enrolled before having this biopsy. The characteristics of these participants were comparable to the total cohort (Table 1). In total, 329 participants (23%) were reclassified at the confirmatory biopsy: 231 (70%) from GG1 to GG2, 76 (23%) from GG1 to GG3 or above, and 22 (7%) from GG2 to GG3 or above.

In the total cohort, the cumulative incidence of biopsy grade reclassification was 43% (95% CI, 40%-45%) at 10 years after diagnosis. In the same time frame, extreme grade reclassification occurred in 14% (95% CI, 13%-16%) of participants and an estimated 49% (95% CI, 47%-52%) of participants underwent treatment (Figure 2; eTable 2 in Supplement 1). In a subcohort analysis, rates were slightly higher in the 1403 participants enrolled before the confirmatory biopsy (Figure 2; eTable 3 in Supplement 1).

In total, 870 participants (40% of total cohort) received treatment, mostly via prostatectomy or radiation (Table 2). There were 425 participants treated after the confirmatory biopsy, and 396 participants treated after a subsequent surveillance biopsy. Treatment in these 2 groups occurred at a median of 1.5 and 4.6 years after diagnosis, respectively. Pretreatment GG and percent of biopsy cores with cancer were similar between those treated after the confirmatory biopsy or after subsequent biopsies (Table 2).

Among the 481 men who underwent prostatectomy, 149 (31%) had adverse pathology (eTable 4 in Supplement 1). The rate of adverse pathology was not apparently different in patients treated after the confirmatory biopsy (31%) or subsequent surveillance biopsies (31%) (Table 2).

The median (IQR) follow-up period after primary treatment was 4.6 (2.3-7.4) years. Of the 823 men treated with prostatectomy or radiation, 76 (9% of treated men; 3.5% of the overall cohort) recurred. The 5-year incidence of recurrence was 11% (95% CI, 7%-15%) in participants treated after the confirmatory biopsy and 8% (95% CI, 5%-11%) in participants treated after subsequent biopsies (eFigure and eTable 5 in Supplement 1).

The majority of participants (>96%) had either not been treated or were treated with no recurrence (Figure 3). In the full cohort, the incidence of developing metastasis 10 years after diagnosis was 1.4% (95% CI, 0.7%-2%), and the 10-year incidence of prostate cancer–specific mortality and overall mortality was 0.1% (95% CI, 0%-0.4%) and 5.1% (95% CI, 3.8%-6.4%), respectively (Figure 2; eTable 2 in Supplement 1). In the subcohort enrolled before the confirmatory biopsy, the 10-year incidence of metastasis, prostate cancer–specific mortality, and overall mortality were 2.1% (95% CI, 0.96%-3.2%), 0.25% (95% CI, 0%-0.75%), and 5.7% (95% CI, 3.8%-7.4%), respectively (Figure 2; eTable 3 in Supplement 1).

Of the 21 participants whose cancer progressed to metastatic prostate cancer, 10 had distant metastasis, 5 had regional metastasis, and 6 had node-positive cancer at the time of prostatectomy (Table 2 and Figure 3). Among the participants with metastasis, 19 (90%) were initially diagnosed with GG1 cancer and 14 (75%) of these participants were treated after confirmatory biopsy, with the majority having reclassification at that time (Figure 3).

Among all participants, 105 died during the study period, including 34 who had undergone cancer treatment. Among those treated, 27 died with no signs of recurrence, 2 had recurrent cancer, and 5 had metastatic cancer, of whom 3 died of prostate cancer (0.1% of all participants).

The goal of active surveillance for prostate cancer is to reduce unnecessary treatments and adverse effects among those diagnosed with favorable-risk cancer and avoid undertreatment of aggressive disease. In this multicenter study in which men with favorable-risk prostate cancer were monitored with protocol-directed PSA measurements and prostate biopsies, the incidence of treatment at 10 years after diagnosis was 49%. Within the same time frame, the incidence of metastasis was 1.4%. The 10-year incidence of prostate cancer–specific mortality was 0.1%, although the overall mortality was 5.1%. The rates of unfavorable outcomes, including adverse pathology at surgical procedure, recurrence after treatment, or metastasis, do not appear to be higher among patients who were treated after several years of surveillance compared to those treated immediately following the confirmatory biopsy. These results are consistent with the premise that surveillance with regular monitoring is a safe management strategy for favorable-risk prostate cancers.

This study included a prospective protocol-driven cohort with high adherence to the biopsy schedules, with 88% of participants undergoing the first follow-up biopsy within 2 years of diagnosis and 97% within 5 years. The high level of biopsy adherence may contribute to lower rates of metastasis than reported in studies that do not require regular biopsies, such as the active monitoring group of the ProtecT study^16,17 and a Canadian population-based cohort.¹⁸ While forgoing biopsies may be appropriate for patients with significant comorbidities or a short life expectancy, adherence to biopsy has been shown to be associated with improved outcomes, such as lower rates of recurrence and metastasis during surveillance.¹⁹

Arguably the most important biopsy after diagnosis is the first biopsy, which occurred at a median (IQR) of 1 (0.7-1.3) year after diagnosis in this study. Some centers will not consider a patient for active surveillance until after this confirmatory biopsy. In contrast, in some settings, a large majority of patients diagnosed with favorable-risk cancer will not have this biopsy.^20,21 In the current study, the rate of grade reclassification at the confirmatory biopsy was 23%, and 30% of reclassifications were to Gleason score GG3 or higher. It is also noteworthy that of the 21 participants who developed metastatic cancer, 13 (62%) were treated shortly after the confirmatory biopsy; all of these participants were initially diagnosed with GG1 cancer, and 10 of the 13 upgraded at confirmatory biopsy. Although it is impossible to determine whether these cases had true biological progression or occult aggressive cancer that was not sampled in the initial diagnostic biopsy, these results highlight the importance of a follow-up biopsy after initial diagnosis of GG1 cancer, a histologic pattern for which the cancer terminology has been challenged.²² Study results also demonstrate that continued monitoring by biopsy identifies cancers that progress in the years following an initial diagnosis of low-grade cancer. Of note, the study protocol calls for a prostate biopsy even with stable serum PSA and normal multiparametric MRI.

Although regular biopsies may contribute to a reduction in metastasis and disease-specific death, study data demonstrate that not all patients receiving active surveillance need to undergo regular prostate biopsies indefinitely. There is clearly a large portion of patients diagnosed with favorable-risk prostate cancer whose cancer does not progress on serial biopsies over 10 to 15 years, and who may be able to avoid biopsies altogether. It is well-established that prostate biopsy is associated with risks of bleeding, infection, urinary retention, and financial toxicity.²³ This group has previously published a risk calculator for patients with GG1 cancer who have a confirmatory biopsy with no reclassification.²⁴ This calculator uses commonly available clinical risk factors such as PSA density (calculated as PSA divided by prostate size) and the number of prior negative biopsy results to identify men who can be safely monitored with a less-intensive surveillance regimen, and the authors are investigating the addition of multimodal strategies, including imaging and biomarkers, that will improve risk stratification with the ultimate goal of personalizing surveillance regimens.

Rates of adverse outcomes after treatment in the men who underwent treatment after several years of surveillance were similar to the rates of adverse outcomes in men treated immediately after confirmatory biopsy. With prolonged follow-up, the data support that delayed treatment does not affect downstream oncologic outcomes and allows for patients to avoid the toxicities of treatment for several years. These results counter the fear of losing the window of curability, providing confidence in the safety of protocol-directed active surveillance.

This large, prospective, multicenter study found that prostate cancers diagnosed as favorable risk and managed with active surveillance have almost entirely benign oncologic outcomes, providing a framework of expectations for clinicians and patients. These findings provide strong support for the use of active surveillance vs immediate treatment of favorable-risk cancer. Furthermore, the findings may help alleviate the fear of PSA screening. Even after the US Preventive Services Task Force changed PSA screening from a D to a C rating,²⁵ there has been a trend of decreased screening and rising metastasis²⁶ and mortality.^27,28 Although a causal relationship has not been established, PSA remains the single most widely utilized and recommended screening test for the early detection of prostate cancer, despite its significant lack of specificity.²⁹ Prostate cancer remains the second leading cause of cancer death among men, with an estimated 34 700 deaths in the US in 2023.²⁷ It is plausible that recent advances in adjunctive biomarkers and MRI³⁰ combined with expanded use of active surveillance could change the benefit-to-harm ratio such that PSA screening could be expanded, improving early detection of aggressive prostate cancers.

This study has limitations. First, an important limitation to note is that study accrual began in 2008 and spanned the years during which multiparametric prostate MRI and biomarker tests beyond PSA began to be used in the diagnosis and management of prostate cancer. The current protocol requires MRI before biopsy, and as MRI came into clinical use, all MRI data have been collected. To date, about half of the cohort has undergone an MRI, including 6 of the participants whose cancer metastasized. It is possible that early use of MRI will further reduce rates of recurrence and metastasis observed in this study. The authors have previously reported on the performance of MRI in surveillance supporting the need for systematic biopsies regardless of MRI results,³¹ a position also supported by recent prostate cancer guidelines.³ Second, enrollment was limited to patients at study sites, and the cohort is therefore primarily White, educated, and insured participants. Although the results may not be generalizable to all populations, especially Black men who are diagnosed with and die of prostate cancer at disproportionately high rates,²⁷ the authors have previously reported that biopsy reclassification was found at similar rates in Black and White men.³² More research on long-term outcomes among Black men on active surveillance is needed. Third, prostate cancer has a long natural history and although the median follow-up in this study of 7.2 years is longer than most published active surveillance cohorts,^5-9 it is likely that with longer follow-up, additional metastases and prostate cancer deaths may occur. These limitations notwithstanding, the PASS cohort represents the largest prospective, protocol-driven surveillance study in North America.

Results from this study demonstrate that 10 years after diagnosis, about half of men on active surveillance for favorable-risk prostate cancer remain free of progression or treatment, less than 2% develop metastatic disease, and less than 1% die of their disease. The findings show that long-term clinical outcomes of adverse pathology, recurrence, or metastasis are equal in men who undergo treatment immediately following the first biopsy after diagnosis when compared with men who undergo treatment after several years of surveillance. The majority of men experience no adverse outcomes from their cancer, offering a clear set of expectations for patients and clinicians to follow.

Accepted for Publication: March 29, 2024.

Published Online: May 30, 2024. doi:10.1001/jama.2024.6695

Corresponding Author: Lisa F. Newcomb, PhD, Fred Hutchinson Cancer Center, 1100 Fairview Ave N, Seattle, WA 98109 (lnewcomb@fredhutch.org).

Author Contributions: Drs Lin and Newcomb had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Newcomb, Zheng, Brooks, McKenney, Tretiakova, Nelson, Lin.

Acquisition, analysis, or interpretation of data: Newcomb, Schenk, Zheng, Liu, Zhu, Brooks, Carroll, Dash, de la Calle, Ellis, Filson, Gleave, Liss, Martin, McKenney, Morgan, Tretiakova, Wagner, Lin.

Drafting of the manuscript: Newcomb, Schenk, Liu, de la Calle, Ellis, Gleave, Martin, Lin.

Critical review of the manuscript for important intellectual content: Zheng, Zhu, Brooks, Carroll, Dash, de la Calle, Ellis, Filson, Liss, McKenney, Morgan, Tretiakova, Wagner, Nelson, Lin.

Statistical analysis: Newcomb, Schenk, Zheng, Liu, Zhu, Gleave, Lin.

Obtained funding: Newcomb, Carroll, Lin.

Administrative, technical, or material support: Newcomb, Brooks, Carroll, Dash, Ellis, Liss, Martin, Tretiakova, Wagner, Nelson, Lin.

Supervision: Newcomb, Zheng, Carroll, de la Calle, Liss, Martin, Morgan, Lin.

Other - pathology review and guidance: McKenney.

Other - patient recruitment: Ellis.

Conflict of Interest Disclosures: Dr Newcomb reported receiving grants from NIH, Canary Foundation, and Institute for Prostate Cancer Research during the conduct of the study. Dr Zheng reported receiving grants from Fred Hutchinson Cancer Center and NIH during the conduct of the study. Mr Zhu reported receiving grants from NIH during the conduct of the study. Dr Liss reported receiving grants from NCI and Canary Foundation during the conduct of the study. Dr Tretiakova reported receiving grants from NIH during the conduct of the study. Dr Wagner reported receiving grants from Canary Foundation during the conduct of the study; being an investor and consultant for PatientApps Inc outside the submitted work; and participating in medical expert witness review for multiple firms. Dr Nelson reported receiving grants from NIH/NCI, Canary Foundation, and Janssen during the conduct of the study; and personal fees from Janssen, Pfizer, and Merck outside the submitted work. No other disclosures were reported.

Funding/Support: This study was supported by Canary Foundation, NIH (U01 CA224255, P30 CA015704, and P50 CA097186) and the Institute for Prostate Cancer Research.

Role of the Funder/Sponsor: Canary Foundation, NIH, and Institute for Prostate Cancer Research had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Data Sharing Statement: See Supplement 2.

Additional Contributions: We thank John L. Gore, MD, MS, University of Washington, for his thoughtful comments on this manuscript. Dr Gore was not compensated for his contributions. We also thank all of the patients and coordinators who have participated in Canary PASS.