Increased Risk for Cancer in Patients with the Peutz-Jeghers Syndrome

Methods

We defined the Peutz-Jeghers syndrome according to Giardiello and colleagues' diagnostic criteria [1]: histopathologically verified hamartomatous polyps with at least two of the following: small-bowel polyposis, mucocutaneous melanotic pigmentation, and a family history of the Peutz-Jeghers syndrome. Thirty-four persons identified through the patient registry of the Mayo Clinic from 1945 to 1996 met these criteria. The registry is a database of all medical diagnoses based on clinical evaluations, surgical procedures, and histopathologic findings. After we identified these patients, we reviewed their medical charts and conducted follow-up telephone interviews. Follow-up information was obtained for all patients. All Mayo Clinic pathology specimens and all available outside specimens were reviewed by one pathologist. Each case of cancer was confirmed by tissue review or pathology reports (one case of cancer was confirmed by verbal report only from outside physician).

To calculate relative risk ratios for cancer, we used SEER (Surveillance, Epidemiology, and End Results) data and age- and sex-specific incidence rates for all types of cancers that occurred between 1986 and 1990 among U.S. white persons [15]. For the calculation of relative risks, a person with more than one type of cancer contributed only the first event to the overall cancer relative risk. Table 1 reports all cancer events subtotals of each type of cancer (for example, gastrointestinal, gynecologic, and breast cancer). For women who had multiple breast and gynecologic cancers, only the first cancer of this type was used to compute the relative risk for gynecologic and breast cancer. By this method, several types of cancer in one person could not inflate the estimated relative risks. Person-years were defined as the number of years that patients were observed from the onset of pigmentation or polyp-related gastrointestinal symptoms to age of cancer onset or age at last contact. The expected number of cancer cases for each organ site was computed by multiplying the age- and sex-specific person-years by the appropriate cancer incidence. The estimated risk ratio was the ratio of the total number of observed to the total number of expected cases of cancer [16]; 95% CIs were based on the Poisson distribution of the observed number of cases of cancer. Two patients were referred to the Mayo Clinic for cancer, not the Peutz-Jeghers syndrome. To minimize referral bias, we did not include these two patients in the relative risk calculations.

Results

Thirty-four patients (from 31 kindreds) who met the criteria for the Peutz-Jeghers syndrome were seen at the Mayo Clinic during the study period [1]. Thirty-one had histopathologic verification of hamartomatous polyps; in the remaining three cases, a relative with the Peutz-Jeghers syndrome had histopathologic verification of polyps at the Mayo Clinic. Eighteen of the 34 patients were male and 16 were female; median follow-up was 20 years. Ninety-one percent of patients had histologic confirmation of hamartomatous polyps, 53% had a family history of the syndrome, and 100% had mucocutaneous pigmentation.

Twenty-six cases of noncutaneous cancer developed in 18 patients (53%). Ten were cases of gastrointestinal cancer, including colorectal (n = 7), duodenal (n = 1), gastric (n = 1), and esophageal (n = 1) cancer. Sixteen cases of extraintestinal cancer occurred in 12 patients. Six patients had multiple primary cancers. One had two primary colon tumors; another initially had colon cancer and then developed prostate cancer. In 1 woman who had had breast cancer, metastatic mucinous adenocarcinoma consistent with colonic or ovarian primary cancer developed several years after diagnosis of breast cancer. On pathology review, the subsequent tumor was found not to be a recurrence of the previous breast cancer. Another woman had bilateral breast cancer, the first case at 30 years of age and the second at 39 years of age. Yet another woman had bilateral breast cancer at diagnosis and simultaneous Dukes A colon cancer. In 1 woman with colon cancer, primary uterine and breast cancers were later diagnosed. Colorectal adenomas were detected in 5 of the 34 patients. One case of colon cancer and the duodenal primary cancer arose within mixed villous and hamartomatous polyps.

In addition to the types of cancer described in the preceding paragraph, rare genital tract neoplasms were found in six women: two mucinous ovarian cystadenofibromas, one Sertoli-Leydig cell stromal tumor, and three sex-cord tumors with annular tubules. The mean patient age at cancer onset was 39.4 years (median, 38.0 years [range, 16 to 59 years]). The mean interval between onset of the Peutz-Jeghers syndrome and diagnosis of cancer was 19.8 years (median, 19.5 years). One woman presented with adenoma malignum of the cervix; she was noted to have buccal pigmentation that had been present since childhood. In all other cases, the Peutz-Jeghers syndrome had been diagnosed before cancer was diagnosed.

During the follow-up period, 53% of patients developed cancer. Twenty-six percent developed gastrointestinal cancer. Fifty percent of women developed gynecologic or breast cancer. Five of these women also had a distinctive, benign genital tract neoplasm; a sixth had a sex-cord tumor with annular tubules but no cancer. Overall, the relative risk for the development of cancer was 9.9 (95% CI, 5.7 to 16.2) (Table 2). For men, the relative risk was 6.2 (CI, 2.5 to 12.8); for women, it was 18.5 (CI, 8.5 to 35.2). In all patients, the relative risk for gastrointestinal cancer was 50.5 (CI, 18.5 to 109.9). This risk was greater in women (150.9 [CI, 31.2 to 440.4]) than in men (30.3 [CI, 6.3 to 88.5]) (P = 0.08). The relative risk for gynecologic or breast cancer was 20.3 (CI, 7.4 to 44.2).

Discussion

In our long-term follow-up study of a cohort of 34 patients with the Peutz-Jeghers syndrome, we confirmed a marked predisposition for both gastrointestinal and extraintestinal cancer. Specifically, we documented primary carcinomas in the colon, duodenum, esophagus, breast, ovary, cervix, uterus, thyroid, lung, and prostate. The relative risk for non-cutaneous cancer was 9.9 (CI, 5.7 to 16.2) in all patients, 18.5 (CI, 8.5 to 35.2) in women, and 6.2 (CI, 2.5 to 12.8) in men (P = 0.001 for the difference between men and women). This difference is explained in part by the excess cases of gynecologic and breast cancer in women (relative risk for gynecologic and breast cancer, 20.3 [CI, 7.4 to 44.2]). In our cohort, however, women also had a significantly increased relative risk for colon cancer. To our knowledge, this disparity in predisposition to cancer between men and women with the Peutz-Jeghers syndrome has not been described previously. Of interest, the distinctive benign neoplasms associated with the Peutz-Jeghers syndrome (sex-cord tumors with annular tubules, sex-cord stromal tumors, and mucinous tumors) typically occur in the female genital tract [8]. In contrast, reports of male genital tract tumors in patients with the Peutz-Jeghers syndrome are infrequent [17]. The mean age of cancer diagnosis in our group, 39.4 years, is young and is entirely consistent with an inherited predisposition.

Giardiello and colleagues [1] observed that colon, pancreatic, stomach, and breast cancer predominated in their study of patients with the Peutz-Jeghers syndrome. Their cohort of 23 patients with histopathologic confirmation of hamartomatous polyps and 8 additional patients whose diagnosis was clinically established had a relative risk for cancer of 18. In another retrospective study of patients with the Peutz-Jeghers syndrome, Spigelman and colleagues [5] described colon, stomach, small intestine, ovary, fallopian tube, thyroid, and lung carcinomas, as well as cancer in which the primary source was unknown. Although our study also has the limitations inherent in a retrospective study, the long-term and complete follow-up of these patients contributes to our understanding of expected outcomes in this rare condition. With the recent linkage and subsequent identification of the gene for the Peutz-Jeghers syndrome on the short arm of chromosome 19 [18-20], genetic testing of persons from Peutz-Jeghers syndrome families may soon become available.

In conclusion, our study confirms that the Peutz-Jeghers syndrome is associated with a significantly increased risk for both intestinal and extraintestinal cancer. Women in affected families seem to have a greater susceptibility to the development of cancer, especially breast and gynecologic cancer. Patients with the Peutz-Jeghers syndrome should receive regular screening for at least colon, breast, and cervical cancer. The screening studies used and the timing of these studies should be tailored to the types of cancer and the age at cancer onset in a given family.