Increased risk of colon cancer after external radiation therapy for prostate cancer

Abstract

Radiotherapy can induce second cancers. Controversies still exist regarding the risk of second malignancies after irradiation for prostate cancer. We evaluated the risk of developing colon and rectum cancers after prostate cancer in irradiated and nonirradiated patients. Using data from the population-based Geneva cancer registry, we included in the study all men with prostate cancer diagnosed between 1980 and 1998 who survived at least 5 years after diagnosis. Of the 1,134 patients, 264 were treated with external radiotherapy. Patients were followed for occurrence of colorectal cancer up to 31 December, 2003. We calculated standardized incidence ratios (SIR) using incidence rates for the general population to obtain the expected cancer incidence. The cohort yielded to 3,798 person-years. At the end of follow-up 19 patients had developed a colorectal cancer. Among irradiated patients the SIR for colorectal cancer was 3.4 (95% confidence intervals [CI] 1.7–6.0). Compared to the general population, the risk was significantly higher for colon cancer (SIR = 4.0, 95% CI: 1.8–7.6), but not for rectal cancer (SIR = 2.0, 95% CI: 0.2–7.2). The risk of colon cancer was increased in the period of 5–9 years after diagnosis (SIR = 4.7, 95% CI: 2.0–9.2). The overall SIR of secondary cancer in patients treated with radiotherapy was 1.35 (p = 0.056). Nonirradiated patients did not have any increased risk of rectal or colon cancer. This study shows a significant increase of colon but not rectum cancer after radiotherapy for prostate cancer. The risk of second cancer after irradiation, although probably small, needs nevertheless to be carefully monitored. © 2008 Wiley-Liss, Inc.

Radiotherapy for cancer, including Hodgkin's lymphoma, cervical, ovarian, testicular, breast and other cancers, has proven to increase the risk of secondary solid cancers, mainly in the radiation field.1–6 Controversy still exists, however, on the risk of colon and rectal cancer after irradiation of the prostate. Some authors found an increased risk of developing colon and rectal cancer7, 8 or rectal cancer only,9, 10 whereas others concluded there was no association between irradiation of the prostate and secondary colorectal cancer.11–13

Using data from the population-based cancer registry, we evaluated the risk of colon and rectum cancers in irradiated and nonirradiated prostate cancer patients in Geneva, Switzerland.

Material and methods

The Geneva Cancer Registry records all incident cancer cases occurring in the population of the canton (approximately 420,000 inhabitants) since 1970. The registry collects information from various sources, and is considered exhaustive, as attested by its low-percentage (<2%) of patients recorded from death certificates only.14 All hospitals, pathology laboratories and practitioners are requested to report all cancer cases. Trained registrars systematically abstract data from medical and laboratory files. Physicians regularly receive questionnaires to secure missing clinical and therapeutic data. Death certificates are consulted systematically. The registry records second primary cancers according to IARC guidelines.15

Recorded data include sociodemographic information, method of discovery, type of confirmation, tumor histology and grade, (coded according to the International Classification of Diseases for Oncology),16 stage of disease at diagnosis, treatment during the first 6 months after diagnosis, survival status and cause of death.

The registry regularly assesses survival, taking as reference date the date of confirmation of diagnosis or the date of hospitalization (if it preceded the diagnosis and was related to the disease). In addition to passive follow-up (standard examination of death certificates and hospital records), active follow-up is performed yearly using the files of the Cantonal Population Office (office in charge of the registration of the resident population). Cause of death is taken from clinical records and coded according to the World Health Organization's classification.17

In this study, we included all men with prostate cancer diagnosed between 1980 and 1998 who survived at least 5 years after diagnosis. We included patients only since 1980, because we had no information about stage of the disease and type of surgery before this date. Patients with any invasive cancer prior to prostate cancer (except skin nonmelanoma), and patients who developed colorectal cancer within 5 years after the prostate cancer, were excluded.

We compared categorical data with χ² analysis. Person-years were calculated from the date of prostate cancer to the date of the colorectal cancer (ICD 9 = 154 or ICD 9 = 153), date of death or December 31, 2003, whichever came first. The expected number of cancer cases was calculated by multiplying the period-age and sex-specific cancer incidence rates for the Geneva population by person-years in 5-year intervals. Cancer incidence rates of the Geneva population for 1980–2003 were used for the calculations. The standardized incidence ratio (SIR) was defined as the ratio between the observed and expected number of cases. A 2-tailed 95% confidence interval (CI) of the SIR was calculated assuming a Poisson distribution of the observed numbers.18 We also calculated the excess absolute risk per 10,000 (EAR) as the difference between the observed and expected number of cases divided by the person-years at risk multiplied by 10,000.

We reopened the medical records of the irradiated patients to extract information regarding radiotherapy regimen, technique, dose in Gray (Gy), fractions, duration, target volume, fields and number of cycles. We also recorded the number and types of complications associated with the radiotherapy treatment.

Finally, to evaluate if there was an effect of dose we calculated by logistic regression the risk of developing colorectal cancer for those patients who received high versus low radiation dose after accounting for other prognostic factors.

All statistical analyses were performed with SPSS software (version 11 SPSS, Chicago, IL).

Results

Of the 1,134 men diagnosed with prostate cancer and surviving at least 5 years, 264 (23%) were treated with primary external radiotherapy. No patient received brachytherapy during the study period. The mean age of the patients was 70 years (range: 44–93). Table I shows the distribution of patients and tumor characteristics for irradiated and nonirradiated prostate cancer patients. Irradiated patients were younger than nonirradiated ones (mean 68 years vs. 71, p < 0.001), the method of detection of their cancer was more often by screening (prostate specific antigen with or without digital rectal examination), the tumor was more often lymph node negative, smaller and better differentiated. Among the 870 nonirradiated patients, 54% (n = 469) underwent prostatectomy compared to only 10% of irradiated patients (Table I). Eleven patients (1.3%) among the nonirradiated received chemotherapy versus none among the irradiated patients.

Table I. Patient and Tumor Characteristics of Irradiated and Nonirradiated Patients with Localized Prostate Cancer

	Irradiated (N = 264)	Nonirradiated (N = 870)	p-Value of χ^2 test
	N (%)	N (%)
1 A cancer is fortuitously discovered during the work-up or treatment of another unrelated illness.— 2 These patients were managed by the “watchful waiting” approach.
Age			<0.001
40–59	28 (11)	88 (10)
60–79	232 (88)	646 (74)
≥80	4 (01)	136 (16)
Period			0.276
1980–1989	92 (35)	284 (33)
1990–1998	172 (65)	586 (67)
Method of discovery			<0.001
Symptoms	133 (50)	357 (41)
Fortuitously1	30 (11)	201 (23)
Screening	96 (36)	232 (27)
Unknown	5 (02)	80 (09)
Socioeconomic status			0.213
High	78 (31)	266 (33)
Medium	96 (38)	343 (42)
Low	75 (30)	190 (23)
Unknown	4 (02)	18 (02)
Lymph nodes			<0.001
Negative	200 (76)	340 (39)
Positive	11 (05)	58 (07)
Unknown	53 (20)	472 (54)
Clinical T stage			<0.001
T0, T1	45 (17)	162 (19)
T2	110 (42)	226 (26)
T3+	93 (35)	287 (33)
Tx	16 (06)	195 (22)
Grade			<0.001
I	81 (31)	210 (24)
II	99 (38)	259 (30)
III	16 (06)	103 (12)
Unknown	68 (26)	298 (34)
Surgery			<0.001
Prostatectomy	79 (30)	469 (54)
Resection	52 (20)	137 (16)
No surgery	133 (50)	2642 (30)
Time since diagnosis (person-years)
5–9 years	741 (79)	2254 (79)
≥10 years	199 (21)	603 (21)

The median follow-up time for the whole cohort was 7.4 years. The irradiated patients were followed for a median of 7.8 years (range 5–22) and the nonirradiated patients for a median of 7.3 years (range 5–24). The person-years at risk of developing a colorectal cancer starting from the 5th year after diagnosis of prostate cancer were 3,798.16 (940.82 and 2,857.34 for those irradiated and those not irradiated, respectively). Overall 19 patients developed a colorectal cancer more than 5 years after a prostate cancer, 6 cancers of the rectum and 13 colon cancers (Fig. 1).

Figure 1. Survival curves for the development of colorectal cancer in the population of men treated with radiotherapy, Geneva 1980–1998.

The risk of colorectal cancer was significantly higher than expected for patients treated with external radiotherapy (SIR = 3.4, 95% CI: 1.7–6.0; EAR = 82.3). The excess risk of colon cancer was 4-fold (SIR = 4.0, 95% CI 1.8–7.6; EAR = 71.6), while the risk of rectal cancer was not significantly increased (SIR = 2.0, 95% CI 0.2–7.2; EAR = 10.6) (Table II).Irradiated patients were more likely to develop colon cancer 5–9 years after diagnosis compared to the general population (SIR = 4.7, 95% CI 2.0–9.2). The risk of colon cancer among irradiated patients remained similarly high, although not significantly, 10 years or more after the prostate cancer. Nonirradiated patients did not have an increased risk of rectal or colon cancer compared to the general population.

Table II. Colorectal Cancer After Prostate Cancer According to Use of External Radiotherapy

Cancer site	Number at risk	Total				5–9 years since diagnosis			≥10 years since diagnosis
		O/E	SIR	95% CI	EAR	O/E	SIR	95% CI	O/E	SIR	95% CI
O, number of observed cases; E, number of expected cases; SIR, standardized incidence ratios; 95% CI, 95% confidence intervals; EAR, excess absolute risk per 10,000 person-years; RT, radiotherapy; * p < 0.05; ** p < 0.01; *** p < 0.001.
Colorectal
RT yes	264	11/3.26	3.4**	1.7–6.0	82.3	9/2.53	3.6*	1.6–6.8	2/0.63	3.2	0.4–11.5
RT no	870	8/10.88	0.7	0.3–1.5	−10.1	8/8.49	0.9	0.4–1.9	0/2.44	–	–
Rectum
RT yes	264	2/1.00	2.0	0.2–7.2	10.6	1/0.81	1.2	0.04–6.9	1/0.1	5.3	0.2–29.3
RT no	870	4/3.33	1.2	0.3–3.1	2.3	4/2.61	1.5	0.4–3.9	0/0.72	–	–
Colon
RT yes	264	9/2.26	4.0*	1.8–7.6	71.6	8/1.72	4.7***	2.0–9.2	1/0.53	1.9	0.1–10.4
RT no	870	4/7.55	0.5	0.1–1.4	−12.4	4/5.88	0.7	0.2–1.7	0/1.72	–	–

Table III shows the characteristics of irradiated patients who developed a colorectal cancer. The second cancer developed on average 8.8 years after the diagnosis of the prostate cancer. The majority of patients received 65/66 Gy to the prostate. Compared to irradiated patients who did not develop colon or rectal cancer, those who did were more often diagnosed and treated in 1980–1989 (64% vs. 34%, p = 0.045), received a dose of ≤67 Gy compared to 68–80 Gy (82% vs. 44%, p = 0.014), and were more likely to have been irradiated with a conventional technique using a 4 field approach followed by a 2 field boost instead of the 6 field conformal technique (55% vs. 17%, p = 0.027). To estimate the adjusted relative risk (RR) of developing a colorectal cancer for the patients who received 68–80 Gy to the prostate compared to patients who received ≤67 Gy of radiations we performed logistic regression models. In a model that included only age and mean dose of radiotherapy the RR for patients treated with 68–80 Gy was 0.2 (95% CI: 0.04–0.91; p = 0.037). In a model adjusted also for socioeconomic status and lymph node (the only 2 other variables statistically significant in the univariate model), the RR of developing a colorectal cancer after mean dose radiations of 68–80 Gy was 0.3 (95% CI: 0.5–1.4; p = 0.112).

Table III. Characteristics of the 11 Prostate Cancer Patients who Developed a Colorectal Cancer After Irradiation

Prostate cancer			Total dose received (Gy)	Primary dose RT(Gy)	Primary fields (N)	Boost dose RT(Gy)	Boost fields (N)	Radiation adverse effects	Colorectal cancer
Age at diagnosis	Stage	Year ofdiagnosis							Year ofdiagnosis	Site
1 The patient interrupted the therapy after 2 treatments.
71	3	1980	64	64	4				1989	Descending colon
72	3	1980	34	34	4				1988	Rectosigmoid junction
75	Unknown	1981	66	50	4	16	2	Dysuria	1988	Ascending colon
73	1	1987	66	50	4	16	2	Diarrhoea	1999	Transverse colon
74	1	1987	41	4^1	4				1994	Transverse colon
78	2	1988	65	45	4	20	2	Pollakiuria	1997	Transverse colon
59	2	1988	66	46		20	2	Dysuria, pollakiuria	2000	Rectosigmoid junction
77	2	1990	65	45	4	20	2		1998	Sigmoid colon
72	3	1990	65	45		20	2	Proctitis, diarrhoea, pollakiuria	1998	Caecum
65	2	1993	68	50	6	18	6	Diarrhoea	2002	Sigmoid colon
73	1	1996	70	54	6	16	6	Rectal bleeding, nocturia	2003	Colon NOS

Overall, irradiated patients had a slight increase of all cancers of borderline significance (Observed = 34, Expected = 25.26; SIR = 1.35; p = 0.056), almost completely due to the increase of colorectal cancers. Of note, there were 3 cases of bladder cancers, of which 2 occurred 10 years after prostate cancer diagnosis vs. 0.39 expected (SIR = 5.15, p = 0.058), and 3 cases of skin melanoma (2 on the trunk and 1 in the face) developed between 5 and 9 years after diagnosis versus 0.74 expected (SIR = 4.08, p = 0.039) (Table IV).

Table IV. Risk of Secondary Cancer in the Population of Men Treatedwith Radiotherapy, Geneva 1980–1998

Type of secondary cancer	Total		5–9 years since diagnosis		≥10 years since diagnosis
	O	SIR	O	SIR	O	SIR
O, number of observed cases; SIR, standardized incidence ratios; * p < 0.05; ** p < 0.01; *** p < 0.001.
All secondary cancers	34	1.35	25	1.28	9	1.55
Buccal cavity pharynx	2	2.32	1	1.48	1	5.35
Tongue	1	7.53
Oropharynx	1	4.05
Digestive System	16	2.42**	14	2.73**	2	1.36
Small intestine	1	8.46	1	10.7
Colon	9	3.99**	8	4.65***	1	1.87
Rectum, RSJ	2	2.00	1	1.23	1	5.26
Liver	2	2.27	2	2.89
Pancreas	1	1.19	1	1.58
Retroperitoneal	1	52.1*	1	66.7*
Respiratory system	3	0.68	2	0.58	1	1.15
Larynx	1	3.22	1	4.05
Lung	2	0.51	1	0.33	1	1.15
Urinary system	4	1.64	2	1.07	2	3.50
Urinary bladder	3	1.84	1	0.80	2	5.15
Kidney	1	1.24	1	1.61
Soft Tissue	1	10.3	1	13.3
Melanoma of skin	3	3.06	3	4.08*
Lymphatic, haematopoietic	4	2.44	2	1.42	2	4.05
Non-Hodgkin lymphoma	1	1.87	1	4.07	1	7.48
Myeloma	1	3.11
Leukemia	2	2.14	1	1.47	1	3.92

Discussion

In this study, we found an overall excess risk of colorectal cancer among irradiated patients, mainly due to an increased risk of colon cancer. These results support the hypothesis of an association between colorectal cancer occurrence and prostate radiotherapy. Using Surveillance, Epidemiology, End Results (SEER) registry data, Brenner et al. reported a risk of rectal cancer increased of 105% and a nonsignificant increase of colon cancer of 24% among patients treated with radiotherapy compared to patients with surgery only 10 or more years after radiation of prostate.10 More recently, reevaluating SEER data, Baxter et al. showed that long-term prostate cancer survivors who received radiotherapy had an increased risk of secondary rectal cancer (hazards ratio = 1.7; 95% CI: 1.4–2.2), compared to nonirradiated prostate cancer survivors. They found no statistically significant association with colon cancer.9 Surprisingly, and in contrast with some previous findings on the same dataset, Mc Master et al. in their last evaluation of secondary cancer risk in SEER data base, found no increased risk for either colon or rectum cancer. However, for cancer of the rectum there is a trend toward an increased risk with increasing time after diagnosis, with a non-significant increase of 13% after 10 years or more.19 On the contrary, Moon et al. found that patients treated with external beam radiation therapy had significantly higher odds of developing both rectum, and sigmoid colon cancer (OR = 1.60; and OR = 1.26, respectively; p < 0.05).8 A population-based Canadian study showed a SIR of 121 for colorectal cancers among patients with prostate cancer who received radiotherapy compared to those without.7

Some of the differences in the results of these studies may be due to differences in the study design and analysis such as study periods, outcome of interest, comparison population, length of follow up, measure of risk, etc. In this population-based study, we included only men who survived at least 5 years after diagnosis to allow time for the development of radiation-induced cancer. The average follow-up of the cohort members is comparable to that of Baxter et al. study.9 In addition, we had information on type, method and dose of radiotherapy treatment.

Rectum and bladder are definitively considered irradiated fields during radiotherapy for prostate cancer. However, other parts of the large intestine can receive substantial amounts of scattered radiation from the volume of the irradiated tissue, particularly the rectosigmoid junction, the caecum, as well as those parts of the large intestine which are “mobile” in the abdomen, like the transverse or sigmoid colon. As a matter of fact, increased colon cancer risk has been observed after pelvic irradiation for female genital benign and malignant conditions, or cancers of the testis.4, 20, 21

Our data did not show a significant increase of rectal cancer risk in the radiation group, although the rectum is heavily irradiated during radiotherapy for the prostate. The limited statistical power of the study may partially account for this result. On the other hand, the nature of the dose-effect dependence for radiation-induced carcinomas is still uncertain, and a reduction in secondary cancer risk at higher doses is also being entertained.22

Elevated risks of solid tumors in patients with prostate cancer may also reflect the influence of diagnostic surveillance, the natural history of the disease itself, and shared etiologic factors. It is, nevertheless, very likely that radiotherapy explains a large part of the observed excess of colon and rectal cancer risk. Among the 11 irradiated diagnosed colorectal cancers, 1 was a stage T1, 1 a stage T2, and all the others were more advanced diseases. This result does not support the hypothesis of anticipated diagnosis as we could expect following a detection bias. We found no increased risk of colon or rectal cancer in nonirradiated patients. Furthermore, we observed the excess incidence rates of colorectal cancer after 8.8 years on average, a reasonable latency period for secondary cancer induction.

This study presents the limitations of cancer registry-based data. Information about cancer therapy comprised only the first 6 months after diagnosis. Also, we lack information on risk factors for colorectal cancer and we do not know if they were more prevalent in prostate cancer patients who received radiotherapy. Furthermore, the size of the study population was small and the period of follow-up relatively short. With a development of a colorectal cancer after an average 8.8 years since diagnosis, the statistical power to find an effect 10 years or more after the diagnosis was very small.

Our results are probably not applicable to the patients treated today. The conventional treatment technique commonly used during the first period of the study involved treatment of the whole pelvis, using a 4 field approach designed to include the prostate, seminal vesicles and the regional lymph nodes for a total of 45–50 Gy. An additional primary target boost was then administered with either a 4 field approach or a bilateral 120° arc rotational technique to increase the dose to the prostate and seminal vesicle.23 With the rotational technique, but even with the 4 field approach, effective shaped blocking cannot be used to shield normal tissues. Thus, a large volume of the bladder and the rectum receives the same dose as the prostatic tumor target.

This, together with the longer follow-up, can explain our observation of a higher proportion of colorectal cancer cases among the patients diagnosed in 1980–1989 and those who received conventional 4 field technique followed by the boost. The observation of a higher proportion of cases among those who received a lower dose of radiation, already seen from Movsas et al.12 is also probably correlated to the technique in use and the volume of exposed nearby organs other than prostate. Today, the use of 3D conformal techniques reduces the effect of radiotherapy on adjacent organs by directing multiple radiation beams at the prostate from several angles.22 The intersection of the beams, thus the region of the highest intensity of radiation, is centered on the prostate.24

In conclusion, we found a significantly increased long-term risk of colorectal cancers after radiation treatment of prostate cancer. This serious long-term side effect should be discussed in view of the benefit of radiotherapy on local tumor control. The future risks of second solid tumours will be probably smaller but need nevertheless a continuous and careful evaluation.

Acknowledgements

The authors wish to thank Ms. Alexandra Ionescu for her help in collecting the data, and Mr. David James for his useful advice and support.