Showing posts with label second primary cancer. Show all posts
Showing posts with label second primary cancer. Show all posts

Tuesday, August 30, 2016

The real risk of secondary malignancies due to prostate radiation

Some of the media (see this Medscape article) have already pounced on a meta-analysis published by a group of researchers from the University of Toronto. The media misguidedly focus on relative risk rather than absolute risk. We examined this very complex subject last year (see this link). Wallis et al. looked at 21 studies in which secondary cancers were reported. They found that men who had external beam prostate radiation also had a higher rate of bladder and colorectal cancers, but not lung cancer or hematological malignancies. This shows association, but not causation.

Incidence of secondary cancers ranges in men who had each therapy were:
  • ·      External beam radiation (EBRT): 0.2-2.3%
  • ·      Brachytherapy (BT): 0.1-2.1%
  • ·      External beam plus brachytherapy boost: 0.2- 1.7%
  • ·      No radiation: 0.3-2.3%


The absolute difference in secondary bladder cancers per 1000 men treated for prostate cancer were:
  • ·      EBRT vs no radiation: +2 (range: -2 to +6)
  • ·      BT vs. no radiation: 0 (range: -4 to +4)


The absolute difference in secondary colorectal cancers per 1000 men treated for prostate cancer were:
  • ·      EBRT vs no radiation: +7 (range: +3 to +10)
  • ·      BT vs. no radiation: 3 (range: -5 to +11)


Risk of secondary bladder cancers was the subject of 9 studies.
  • ·      7 of the 9 did not find significantly increased risk
  • ·      The average increase was 39%
  • ·      Of the 3 studies that allowed for a 5 year lag time for secondary cancers to develop after radiation exposure,:

o   2 of the 3 did not find significantly increased risk
o   The average increase was 30%
  • ·      Of the 2 studies that allowed for a 10 year lag time for secondary cancers to develop after radiation exposure,:

o   1 of the 2 did not find significantly increased risk
o   The average increase was 89%
  • ·      Of the 4 studies that included a multivariate analysis that included age:

o   2 of the 4 did not find significantly increased risk
o   The average increase was 67%

Risk of secondary colorectal cancers was the subject of 10 studies.
  • ·      5 of the 10 did not find significantly increased risk
  • ·      The average increase was 68%
  • ·      Of the 4 studies that allowed for a 5 year lag time for secondary cancers to develop after radiation exposure,:

o   2 of the 4 did not find significantly increased risk
o   The average increase was 94%
  • ·      Of the 2 studies that allowed for a 10 year lag time for secondary cancers to develop after radiation exposure,:

o   1 of the 2 did not find significantly increased risk
o   The average increase was 56%
  • ·      Of the 3 studies that included a multivariate analysis that included age:

o   1 of the 3 did not find significantly increased risk
o   The average increase was 79%

While the media focus on the increase in incidence when expressed as a percent, the actual incidence and the absolute increases are really very small.  This was the subject of an accompanying editorial by Eyler and Zietman. They also point out that many of the studies included in the meta-analysis included men treated with older forms of external beam radiation that lack the accuracy of today’s technology. We expect to report soon on an analysis of IMRT only.

Many of the studies lacked data on age and other risk factors that contribute to bladder and colorectal cancers. Patients who received EBRT are an average of 10 years older than patients receiving surgery and about 5 years older than patients receiving BT. It is established that bladder and colorectal cancer incidence increase with age, so it is difficult to separate the competing risk factors. The problem is compounded by the lag time necessary to observe secondary cancers. Other risk factors confounding any such analysis include smoking and ethnicity. Because we are increasingly vigilant after a first cancer diagnosis, we are more likely to detect other cancers. This also confounds our analyses.


All of these studies really only tell us about association but not causation. For that, we require randomized clinical trials with very long tracking. While such trials are a long way off, we can take comfort in the fact that the risk is really very small.  While any risk should be acknowledged, and may be a decision factor for choosing active surveillance in low-risk men, this should not be a reason for anyone to avoid needed and curative radiation therapy.

Saturday, August 27, 2016

Radiotherapy-induced second primary cancer (RTSPC) risk is low and may differ by radiation technique.


While radiation therapy cures many cancers, many men worry that radiation-induced sub-lethal damage to healthy tissue may someday result in a second primary cancer in the irradiated field. Several studies have sought to quantify that risk, but it is difficult to quantify for several reasons:
  • ·      Studies have to follow patients for a very extended length of time before radiation-induced second cancers are likely to show up. Based on survivors studied in Hiroshima and Nagasaki, five years is the minimum, and most will show up by 15 years.
  • ·      An attempt must be made to separate second primary cancers due to other causes from second primary cancers due to radiation.
  • ·      There are many confounders, especially age, race, smoking, lifestyle, family risk and genetic variants. Also, there is an inherent detection bias because patients are already monitored for their first primary prostate cancer, increasing the likelihood that a second primary cancer will be detected.
  • ·      The patient base for radiation therapy has changed over the years. Until the mid 1990s radiation was often only given to men with lower life expectancies. Therefore, many second malignancies never showed up in the databases.
  • ·      There has never been a study, nor is there likely to ever be one, where patients are randomized to receive radiation or receive no radiation and are watched for 10+ years.
  • ·      The risk is quite small, so data from a single institution will never have the power to find statistically significant effects.
  • ·      Because of the above, the only useful data must come from large databases, and an attempt must be made to match the samples or otherwise control for other variables that affect outcomes. Such databases don’t always have the information needed for proper matching and controls.
  • ·      Over the years, there have been a wide variety of radiation technologies and doses used. By the time the data has matured sufficiently to be used, the results are irrelevant because the technology is obsolescent. 

In spite of those difficulties, there have been several attempts to quantify the risk. Without getting into the details, I’ll just mention a few. Abdel-Wahab et al. (2008) found an extra 1.6 radiotherapy-induced second primary cancers (RTSPCs) per 1,000 men (0.16%) receiving external beam radiation therapy for prostate cancer compared to men with prostate cancer who had no surgery or radiation in the period of 1972-2002. They found no difference between brachytherapy and external beam radiation. In another study by Abdel-Wahab et al. (2009), they found a higher rate – 3.7 RTSPCs per 1,000 - in patients receiving salvage radiation for a recurrence after primary prostatectomy. This is not surprising because salvage radiation increases the volume of healthy tissue exposed to radiation. Black et al. (2013) found that radiation therapy increased the RTSPC rate by 4.1 per 1,000. A recent meta-analysis by Murray et al.(2014) concludes the RTSPC rate is on the order of 3-5 per 1,000. Berrington de Gonzalez et al. (2011), looking at all cancers (not just prostate cancer), calculated that there were 5 excess cancers per 1,000 patients treated by radiotherapy by 15 years after diagnosis. The risk seems consistent and small in magnitude across studies.

In a study of 2,658 patients treated at Memorial Sloan Kettering between 1998 and 2001, Zelefsky et al. (2012) found that after 10 years of follow-up, there were no extra cancers attributable to external beam radiation or brachytherapy. All extra cancers were completely explained by age and smoking. Hamilton et al. (2014) looked at 2,418 patients treated with brachytherapy and 4,015 treated with prostatectomy at the British Columbia Cancer Agency between 1998 and 2010. There was no difference in pelvic second primary cancer rates between the two groups. As in the MSKCC study, all differences were attributable to age and smoking. It’s likely that these cohorts have to be at least ten times larger to observe any effect. However, these two studies that looked at more recent treatment data raises the question as to whether the already small risk of RTSPC has diminished still further with improved radiation technology.

I thought it important to establish the small size of the absolute risk as a prelude to a discussion of a comparative risk study.  Berrington de Gonzalez et al. (2015) looked at patients in the SEER-Medicare database from 1992 to 2004 who were 66 to 84 years of age and diagnosed with prostate cancer, and followed them through 2009. They filtered out secondary cancers appearing sooner than 5 years from diagnosis (2 years for leukemias), and adjusted for age, length of follow-up, co-morbidities, hormone therapy and chemotherapy. There were 40,000 patients in the cohort. They looked at the following radiotherapy parameters/technologies:
  1.  2DRT or 3DCRT: 3-dimensional conformal radiation therapy (3DCRT) came into vogue in the mid-1990s and largely replaced 2-dimensional radiation therapy (2DRT). 2DRT featured two opposed radiation fields aimed at the target and filtered only by lead blocks. 3DCRT featured radiation fields with beams aimed from many angles and shaped to the topography of the target with multi-leaf collimators. The beams are adjusted according to the 3D shape found on a planning CT scan. It gives much lower doses to nearby organs. IMRT was not widely adopted until about 2002, but has rapidly replaced 3DCRT for most prostate cancer treatments. So for most of us, the findings are of historical interest only.
  2.  The energy of the X-rays used (>10 MV or ≤10 MV): As 3DCRT supplanted 2DRT, and linear accelerators replaced Cobalt-60 machines, higher energy beams could be used that could penetrate deeper with diminished damage to surrounding organs.
  3. Brachytherapy: This became popular in the 21st century
  4. External beam radiation with a brachytherapy boost to the prostate: Also, not very much used until the 21st century.
The authors found:
  • ·      Secondary rectal cancers were reduced by 41% among patients who were treated with 3DCRT compared to 2DRT.
  • ·      There were no differences in the relative risk of any kind of cancer except rectal cancer, according to type of therapy or energy used. It was perhaps surprising that there was no increase in bladder cancer or colon cancer.
  • ·      Comparing brachytherapy to external beam radiation, the relative risks of secondary cancers were lower (but not significantly so, except for colon cancer and leukemia) for brachytherapy in the first 5 years of follow-up, but were the same or higher (but not significantly so) during the next six years of follow-up for every cancer type. This is probably due to an early detection bias rather than a real effect.
Dr. Berrington de Gonzalez is continuing to track the SEER-Medicare database and hopes to analyze the effects of IMRT later this year. In the meantime, we can only look to modeling studies for clues about contemporary radiotherapy techniques.

Murray et al.(2015) simulated 6 different radiation techniques in three patients to determine which delivered the lowest organ-equivalent doses to organs at risk for secondary cancer development. The six techniques they studied were:
  • 1.     3DCRT (using 10 MV X-rays) – 78 Gy across 39 treatments (see explanation above).
  • 2.     Intensity Modulated Radiation Therapy (IMRT) (using 6 MV X-rays from 5 directions) – 78 Gy across 39 treatments: This is the most prevalent kind of radiation used for prostate cancer. IMRT not only conforms to the shape of the prostate, like 3DCRT, but also uses beams of varying intensity and of lower energy so that many fewer of them penetrate beyond the prostate.
  • 3.     Volumetric Modulated Arc Therapy (VMAT) – 78 Gy across 39 treatments: VMAT is similar to IMRT except that the radiation is applied continuously in an arc around the patient.  IMRT is often called “step and shoot,” which takes considerably longer per treatment. VMAT’s faster speed of treatment allows for less organ movement within the treatment, and therefore, less potential “overspray” to organs at risk.
  • 4.     Stereotactic Ablative Body Radiotherapy (SABR), also known as Stereotactic Body Radiation Therapy (SBRT) – 42.7 Gy across 7 treatments: SABR applies the radiation in fewer treatments or fractions (called hypofractionation). Because prostate cancer is particularly sensitive to killing with higher dose per fraction, lower total doses are needed. This delivers lower doses to organs at risk. It can be delivered with a variety of linear accelerators. In this study, SABR was delivered using the VMAT machine.
  • 5.     Flattening Filter-Free (FFF): A flattening filter changes the shape of each beam so that it is flat rather than peaking in the center. This slows the treatment somewhat, and is of doubtful value in prostate treatment. It is not used with IMRT because the multi-leaf collimator makes it redundant. In this study, the researchers tried both of the VMAT plans (the normally fractionated one  (#3) and the hypofractionated SABR one (#4)) with the filter in place and without it.


The authors used several different radiobiological models to estimate the “excess absolute risk” (EAR) of a second cancer per 10,000 person-years. They studied the second-cancer risk for organs within the radiation field (rectum, bladder, pelvic bones and soft tissues) as well as out-of-field organs exposed to low scattered doses. The key findings of the modeling study were:

  • ·      The EAR for the 3DCRT plan was about 2.7 excess second cancers per 10,000 persons-years for rectal cancer, about 2.3 for bladder cancer, about 0.8 for the colon, and about 0.1 for pelvic bone or soft tissue cancers.
  • ·      The 3DCRT, IMRT, and VMAT plans all had about the same relative risk of RTSPCs of the bladder, rectum and colon. Compared to 3DCRT, the SABR plans had about half the relative risk of RTSPCs of the rectum and colon, and about three-quarters the relative risk of RTSPCs of the bladder.
  • ·      There were no significant differences attributable to the use of the flattening-free filter on in-field organs. The FFF plans only made a significant difference to the most distant organs that had negligible risk anyway.


There are some differences between the researchers’ use of SABR in their study and the way it is typically used in the U.S.:
  • ·      They used a 6 mm margin around the prostate, whereas margins half that size or less are commonly used in practice in the U.S.
  • ·      They assumed that image-guidance techniques would only be used to account for prostate movement between treatments, whereas U.S. practice usually entails tracking movement during the treatment as well.
  • ·      The degree of hypofractionation they used (42.7 Gy across 7 treatments) is not as extreme as in typical U.S. practice (35-40 Gy across 5 treatments).

These differences would lower the RTSPC risk of SABR in the U.S. even further.

The major conclusion is that the RTSPC risk is quite low regardless of which of the six radiation therapy technologies were used. Of the technologies, SABR has the lowest potential RTSPC risk.

note: Thanks to Dr. Amy Berrington-Gonzalez for providing access to the full text of her NCI-sponsored study.