Assessment of Second Primary Cancer Risk Among Men Receiving Primary Radiotherapy vs Surgery for the Treatment of Prostate Cancer

Key Points Question Do patients with prostate cancer in the modern era have a higher incidence and risk of developing a second primary cancer after radiotherapy compared with nonradiotherapy treatments? Findings In this cohort study of 143 886 patients with localized prostate cancer, receipt of radiotherapy was associated with a small but statistically significant increase in the risk of developing a second primary cancer. Meaning This study’s findings suggest that it is important that shared decision-making among patients with prostate cancer includes discussion of late toxic effects, including the risk of developing a second primary cancer, and that the possibility of developing a second primary cancer after radiotherapy be considered during evaluation of symptoms among survivors.


Introduction
Localized prostate cancer is a distinctly preference-sensitive condition because radical prostatectomy, radiotherapy, and, in some cases, active surveillance have equivalent long-term cancer-specific outcomes. 1The risk profiles of each treatment option differ; therefore, shared decision-making is an important part of the prostate cancer treatment approach. 1Patients need to feel confident in understanding how the risks and benefits pertain to them.Given that all treatment options are associated with excellent prostate cancer-specific mortality rates, 2 discussion of the potential adverse effects of surgical procedures and radiotherapy is especially important to help patients select treatment that best suits their goals.
The risks associated with radical prostatectomy are well understood because of the availability of many surgical outcomes data sets that cover the brief perioperative period, during which complications become apparent.During radiotherapy, clinicians similarly assess patients for the presence of acute toxic effects and discuss symptom management.However, late toxic effects can present months to years after completion of radiotherapy when patients may no longer regularly receive care from a radiation oncologist.
One late toxic effect, radiotherapy-induced cancer, specifically sarcoma occurring in previously irradiated bone, was first described in the 1940s. 3The original definition of radiotherapy-induced cancer, still in use today, is a cancer that (1) occurs in an irradiated field, (2) has a latency period of at least 4 years between completion of radiotherapy and diagnosis of the new cancer, (3) has different histological features than the primary cancer, and (4) was not present before radiotherapy exposure. 3,4If all required elements of this definition cannot be confirmed, the term second primary cancer is used to describe cancer diagnosed in patients who previously received radiotherapy.In this study, we were not able to unequivocally confirm part 4 of the definition of a radiotherapy-induced cancer; therefore, we referred to a cancer diagnosed in a previously irradiated field as a second primary cancer.
Prostate radiotherapy exposes the bladder, rectum, and other nearby structures, including bone marrow, to high doses of radiation; however, data regarding second primary cancers in these organs are conflicting.7][8] An older study found an increased risk of second primary rectal cancers that was higher than the background rate, 9 whereas more current data 10,11 have suggested this risk is no higher than that found in the general population.Hematologic cancers historically have not been associated with prostate radiotherapy 5 ; however, they have been associated with pelvic radiotherapy for the treatment of other cancers. 6 update the assessment of long-term risk of developing a second primary cancer after receipt of primary radiotherapy for the treatment of localized prostate cancer, we examined patients diagnosed with prostate cancer from January 1, 2000, to December 31, 2015, who were receiving care from the Veterans Affairs (VA) health care system.Data were analyzed from May 1, 2021, to May 22, 2022.We hypothesized that the incidence and risk of developing a second primary cancer within the relevant anatomical field would be higher among patients who received primary radiotherapy compared with those who received nonradiotherapy treatments, including surgical procedures, active surveillance, medical management, or observation.

Data Source
We conducted a retrospective analysis of the incidence and risk of developing a second primary cancer after treatment for prostate cancer using data from the VA Corporate Data Warehouse (CDW).
The Stanford Institutional Review Board approved this study.A waiver of informed consent was granted because the retrospective study design posed minimal risk to participants.This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cohort studies.
We identified patients and captured clinical and treatment characteristics in the CDW oncology data set, which includes cancer diagnoses, treatment information, and clinical characteristics of patients nationally from VA medical centers that diagnose or treat cancer.We supplemented clinical information from the CDW oncology data set with patient Veterans Health Administration electronic health record information using the Observational Medical Outcomes Partnership Common Data Model, 12 which provides standardized terms for clinical data to facilitate research across multiple systems and periods.

Study Cohort
We identified men We compared patients who received primary radiotherapy in the year after diagnosis (radiotherapy cohort) with their counterparts who received a primary surgical procedure, active surveillance, medical management, or observation (nonradiotherapy cohort).

Outcomes
Our

Statistical Analysis
We compared demographic and clinical characteristics between patients in the radiotherapy and nonradiotherapy treatment cohorts using χ 2 tests for categorical variables and Wilcoxon rank sum tests for continuous variables.We assessed unadjusted survival in each treatment group using a life table.We used Cox regression analysis to estimate the association between radiotherapy and the risk of a second primary cancer, adjusted for all covariates of interest.To test the proportional hazard assumption, we included an interaction term between treatment and time.The interaction was statistically significant, suggesting the treatment effect changed over time; therefore, we calculated a piecewise model with radiotherapy effect assessed at 5-year intervals.Time to event was defined as time from the diagnosis of prostate cancer to the diagnosis of a second primary cancer.Patients without a second primary cancer were censored at the date of their last VA visit or the date of death through December 31, 2020.
To compare men who were eligible for active therapy, we performed a sensitivity analysis comparing patients in the radiotherapy cohort with patients in the nonradiotherapy cohort who received a surgical procedure (excluding patients who received active surveillance, medical management, or observation because these patients may have had more medical comorbidities than those eligible for active treatment) using a piecewise Cox model adjusted for the same covariates included in the main analysis.We performed 3 additional sensitivity analyses to confirm no difference in the primary end point.

Results
Among 143 886 male veterans with localized prostate cancer who met inclusion criteria (Figure 1 2).The number needed to harm, defined as the number of patients needed to receive radiotherapy for 1 additional patient to develop a second primary cancer, decreased with the amount of time from

Discussion
In this large national cohort study of men with localized prostate cancer who were followed up for a median of 9 years, treatment with primary radiotherapy was associated with a higher risk of developing a second primary cancer than treatment without radiotherapy.Although the higher incidence and risk of developing a second primary cancer were relatively small (occurring in only 3.0% of patients), the risk increased over time after completion of radiotherapy, and the number needed to harm notably decreased at 10 years after treatment.Moreover, prostate cancer is common, occurring in approximately 1 in 8 men in the US, and even a low-risk event can have consequences for a large absolute number of patients.Careful consideration of risks and benefits among accepted treatment modalities is therefore paramount.Our findings confirmed and updated conventional wisdom pertaining to a specific long-term risk of radiotherapy (especially Ն10 years after treatment) and supported consideration of the use of surgical procedures among otherwise healthy patients diagnosed in earlier decades of life compared with older patients or those with higher comorbid disease burden.
To our knowledge, this study of more than 100 000 patients with prostate cancer represents the largest cohort study to date assessing second primary cancers after prostate radiotherapy.The study cohort received treatment in the modern era of radiotherapy, whereas much of the previous literature, as reported in a systematic review, 16 has focused on patients who received treatment from the 1970s to the early 2000s.During the 1970s to 1990s, patients were more likely to receive treatment with older radiotherapy techniques, such as 2-dimensional or 3-dimensional conformal radiotherapy, which are less conformal than techniques typically used today, such as intensitymodulated radiotherapy, volumetric arc radiotherapy, and stereotactic body radiotherapy.Although the CDW oncology data set lacks information regarding radiotherapy doses, fields, and techniques, our study period was one during which intensity-modulated radiotherapy, volumetric arc radiotherapy, and stereotactic body radiotherapy had been accepted practice standards for prostate radiotherapy. 17These techniques generally reduce the high radiation dose to normal tissues but may increase the low dose to other tissues, with some potential risk. 18e current study specifically compared a cohort of patients with prostate cancer who received primary radiotherapy with a cohort who did not receive radiotherapy, an important distinction from previous literature 16,[19][20][21][22][23] comparing patients with prostate cancer who received radiotherapy with the general population.Our findings in the modern radiotherapy era are supported by older data comparing patients with prostate cancer who received radiotherapy with those who did not, which also revealed an increased risk of bladder and rectal cancers after radiotherapy. 5,9,16,20,21Using the general population as a comparison group may be inappropriate because patients with prostate cancer are treated by urologists, who are more likely than primary care physicians to identify signs and symptoms of bladder cancer, for example.7][8][9][10][11] In contrast to previous studies, 11,20,21 bladder cancer was the most frequently diagnosed second primary cancer in the present study cohort, which may reflect the use of data from the VA population, among whom bladder cancer is the fourth most frequently diagnosed cancer. 24We also found that leukemias and lymphomas, which are not commonly addressed in the literature examining second primary cancers after prostate radiotherapy, were diagnosed more frequently than rectal cancer.Pelvic radiotherapy exposes the bone marrow to potentially high doses, and patients receiving prostate radiotherapy have been found to have an increased incidence of anemia. 25Radiation exposure is a risk factor for lymphoma, 26 and a retrospective study 18 of patients with prostate cancer who received treatment from 1999 to 2011 reported an increased risk of myelodysplastic syndrome and acute myeloid leukemia among patients receiving radiotherapy vs a surgical procedure (HR, 1.51).In that study, 18 which used data from the Surveillance, Epidemiology, and End Results and Medicare databases, the median time to development of myelodysplastic syndrome or acute myeloid leukemia was approximately 3 years for both groups.Given the definition of radiotherapy-induced cancer, it is unclear whether these increased risks were associated with radiotherapy or whether the definition of radiotherapy-induced cancer needs to be updated.Our data contribute to the literature by revealing that primary radiotherapy for the treatment of prostate cancer was associated with a small but significantly higher risk of both leukemias and lymphomas compared with other treatment modalities.It is important for clinicians to be aware of this risk and follow up patients accordingly.

Limitations
This study has limitations.Although we were able to classify patients as receiving vs not receiving radiotherapy, it is plausible that some misclassification occurred because patients who did not receive radiotherapy may have sought treatment outside the VA health care system.However, the VA Cancer Registrars capture initial prostate cancer treatments delivered outside of the VA health care system.In addition, some patients in the nonradiotherapy cohort may have received adjuvant or salvage radiotherapy in the years after their initial diagnosis, which may not have been captured.As with any retrospective cohort study, our data carry a risk of selection bias, and patients treated with radiotherapy may have had more comorbid disease with a higher risk of developing a second primary cancer.Although we were able to adjust for recorded comorbid disease counts using the wellestablished Prostate Cancer Comorbidity Index, it is possible that some comorbid diseases were not recorded in the patient records.We did not capture all potential risks associated with radiotherapy or surgical procedures, including outcomes that are difficult to ascertain using medical records but highly relevant to patients, such as sexual and urinary dysfunction.Although we performed a sensitivity analysis that included a smoking history covariate, information on pack-year smoking history was unreliable, and this variable was therefore not included as a confounder for a second primary bladder cancer.

Conclusions
In this cohort study, patients with prostate cancer who received primary radiotherapy were more likely to develop a second primary cancer than those who did not receive radiotherapy, with increasing risk over time, especially 10 years or more after treatment.Overall, the incidence and risk of developing a second primary cancer were relatively low in both groups.Although the long-term toxic effects of radiotherapy are important to discuss when counseling patients on the risk-benefit profile of prostate radiotherapy, they need not deter physicians from recommending radiotherapy if appropriate.Patient selection and shared decision-making remain important when considering prostate cancer treatment options.Physicians caring for patients who have previously received The median (IQR) time to development of any second primary cancer was 6(3-9) years after the prostate cancer diagnosis (median [IQR], 6.2 [3.5-9.5] years in the radiotherapy cohort and 5.5 [3.2-8.7] years in the nonradiotherapy cohort).The median (IQR) time to development of the 4 most frequent second primary cancers in the radiotherapy cohort was 6.2 (3.5-9.4) years for bladder cancer, 5.8 (3.6-9.2) years for leukemia, 5.7 (3.6-8.7) years for lymphoma, and 6.3 (3.3-9.9) years for rectal cancer; in the nonradiotherapy cohort, the median (IQR) time was 5.2 (3.1-8.4) years for bladder cancer, 5.7 (3.3-8.9) years for leukemia, 5.9 (3.3-9.0) years for lymphoma, and 5.0 (2.8-7.9) years for rectal cancer.

Downloaded From: https://jamanetwork.com/ on 09/16/2023 spread
15 the seminal vesicles, and T4 indicating the tumor has spread to tissues next to the prostate other than the seminal vesicles); and score on the Prostate Cancer Comorbidity Index 14 (a weighted scale based on age and comorbidity that was validated among men with prostate cancer to estimate long-term noncancer mortality; scores were categorized as 0, 1-2, 3-4, and Ն5, with higher scores indicating older age and greater number of comorbidities), which was calculated using electronic health record data.Missing data on PSA values or Agent Orange exposure in the CDW oncology data set were supplemented by searching the health record for the most recent PSA value in the 6 months before diagnosis or any indication of Agent Orange exposure.A missing category was added for variables with incomplete data.A substantial amount of data on smoking status at diagnosis was missing in early diagnosis years (eg, 66% missing in 2000); smoking status was therefore incorporated only in a sensitivity analysis.We categorized smoking status into never, former, and current based on the most frequently occurring response before diagnosis, as previously described.15 similar between the cohorts (66[61-71]years in the radiotherapy cohort vs 65 [60-72] years in the nonradiotherapy cohort).Patients in the radiotherapy cohort vs the nonradiotherapy cohort had higher Gleason scores (eg, Gleason score >8 points: 8601 patients [16.3%] vs 8944 patients [9.8%]; P < .001)and higher clinical tumor stages (eg, stages T2 and T3: 17 089 patients [32.3%] vs 25 837 patients [28.4%];P < .001).A greater proportion of patients in the radiotherapy cohort were Black or 52 886 patients (36.8%) received primary radiotherapy, and 91 000 (63.2%) did not receive primary radiotherapy.Among those who did not receive radiotherapy, 31 218 patients (34.3%) received a surgical procedure, and 59 782 patients (65.7%) received active surveillance, medical management, or observation.Demographic and clinical characteristics of the radiotherapy and nonradiotherapy cohorts at the time of prostate cancer diagnosis are shown in Table 1.The median (IQR) age was JAMA Network Open | Oncology Second Primary Cancer After Radiotherapy vs Surgery for Prostate Cancer JAMA Network Open.2022;5(7):e2223025.doi:10.1001/jamanetworkopen.2022.23025(Reprinted) July 28, 2022 4/12 Downloaded From: https://jamanetwork.com/ on 09/16/2023

Table 1 .
Participant Characteristics (continued) Likelihood of prostate cancer recurrence based on Gleason score, PSA value, and clinical tumor stage.Range, 1-10, with higher scores indicating greater likelihood of cancer growth and spread.T1 indicates the tumor cannot be felt during a direct rectal examination or observed on imaging but may be detected when a surgical procedure is performed for another condition, T2 indicates the tumor appears to be confined to the prostate, and T3 indicates the tumor has grown outside the prostate and may have spread to the seminal vesicles but has not spread to other tissues next to the prostate.< .001];years 10-15: HR, 1.58 [95% CI, 1.31-1.90;P < .001];years 15-20: HR, 1.46 [95% CI, 0.93-2.29;P = .10]).(eTable 2 in the Supplement).

Table 2 .
Second Primary Cancer-Free Survival and Number Needed to Harm After Prostate Cancer Diagnosis Among Radiotherapy and Nonradiotherapy Cohorts Downloaded From: https://jamanetwork.com/ on 09/16/2023