Cancer risk in patients with Peutz–Jeghers syndrome: A retrospective cohort study of 336 cases

Abstract

Peutz–Jeghers syndrome is a rare autosomal dominant inherited disorder characterized by mucocutaneous pigmentation and hamartomatous gastrointestinal polyposis. A growing body of evidence has shown that Peutz–Jeghers syndrome could cause an increased risk of various cancers, yet the range of cancer risk estimates was wide among different studies. In this retrospective cohort study, 336 patients with Peutz–Jeghers syndrome in China were enrolled. The clinical characteristics, cancer spectrum, relative cancer risks, and cumulative cancer risks were analyzed. In total, 52 patients were diagnosed of cancer in the follow-up period, at a median age of 41 years (range: 21–67). The relative risk for cancer in Peutz–Jeghers syndrome patients was 63.858 (confidence interval: 47.514–85.823), and the cumulative cancer risk at the age of 60 years was 55%. Colorectal cancer was the most common cancer for Peutz–Jeghers syndrome patients (relative risk: 237.918, confidence interval: 154.417–366.572) and the cumulative cancer risk at the age of 60 years was 28%. There was a statistically significant difference in the cumulative cancer risk between patients with family history and those without family history, as well as between patients living in rural area and those living in urban areas (p < 0.05), while no significant effects of gender and intussusception history on the cumulative cancer risk was found (p > 0.05). Hopefully, our study may contribute to the management of this rare disorder and establishment of related surveillance projects, especially in China.

Introduction

Peutz–Jeghers syndrome (PJS), named after the physicians Peutz and Jeghers for their recognition and description of this disease in the early 20th century, is a rare autosomal dominant inherited disorder.^1,2 It is characterized by gastrointestinal hamartomas, mucocutaneous pigmentations, and heredity. The estimated incidence of the disease was between 1: 8300 and 1: 200,000 live births.^1,3 Germline mutation in the serine/threonine kinase 11 gene (STK11, also known as liver kinase B1 (LKB1)) has been reported to cause PJS,^4,5 which can be found in 75%–94% of PJS patients with currently available techniques.^6,7 It remains debatable whether a second gene locus exists or not.⁸

Besides the complications, such as anemia, bleeding, abdominal pain, and intussusception caused by hamartomas at a young age,⁹ there is a growing body of evidence that has shown that PJS also causes an increased risk of cancers, including gastrointestinal cancers and extragastrointestinal cancers.^6,10,11 Among gastrointestinal cancers, increased cancer risk was indicated for the colon, the stomach, the small intestine, and the pancreas. While among extragastrointestinal cancers, gynecological cancers including ovarian cancer, cervical cancer, uterine cancer, and breast cancer were of great risks for female PJS patients. Meanwhile, high risk of pulmonary cancer, renal cancer, prostatic cancer, bone cancer, and leukemia has also been reported.^6,10,11 Due to the rareness of PJS and lack of enough large cohort studies, there is a wide variability in cancer risk estimates, as reviewed in a recent meta-analysis study.¹²

It is clinically important to estimate cancer risks of PJS patients in the implementation of cancer surveillance programs and development of chemopreventative strategies.¹³ Since PJS is an autosomal dominant inherited disorder and cancer is the outcome of gene–environment interactions,¹⁴ the genetic polymorphism among different races,¹⁵ as well as the different lifestyles among different countries, should be known.¹⁶ However, to our knowledge, despite several case reports,^17–19 there was no first-hand retrospective cohort studies reported in China. In this study, 336 PJS patients were enrolled, with cancer being diagnosed in 52 patients. The clinical characteristics, cancer spectrum, relative cancer risks, and cumulative cancer risks of these PJS patients were analyzed.

Materials and methods

Participants

The study was designed as a retrospective cohort study. A total of 336 PJS patients ascertained in General Hospital of Air Force, from 2 March 2005 to 3 May 2016, were enrolled in this study. All the patients included in this study were of Han nationality (in fact, most of the Chinese people are of Han nationality). The entire study was approved by the ethics committee of General Hospital of Air Force, People’s Liberation Army (PLA), China. All patients signed an informed consent to participate in this study. The clinical data were obtained by interview (including telephone interview) and chart review.

Clinical diagnosis was made according to the clinical criteria, as previously reported,¹¹ that a patient will be considered being affected in presence of any one of the following: (1) two or more histologically confirmed Peutz–Jeghers polyps, (2) any number of Peutz–Jeghers polyps detected in one individual who has a family history of PJS in close relative(s), (3) characteristic mucocutaneous pigmentation in an individual who has a family history of PJS in close relative(s), and (4) any number of Peutz–Jeghers polyps in an individual who also has characteristic mucocutaneous pigmentation.

Statistical analysis

Statistical analysis was performed with SPSS 19.0 (IBM, Armonk, NY, USA). Relative risks (RRs) for cancers were calculated by comparing with the general population in China (Chinese cancer registry annual report 2012) as standard.²⁰ Kaplan–Meier method was employed for analyzing the cumulative cancer risks (p < 0.05 was considered to be significant).

Results

Clinical characteristics

A total of 336 PJS patients were enrolled in this cohort. At the closing date of the study, 52 patients (15.5%) were diagnosed of cancer. Baseline characteristics were compared between patients with cancer and those without cancer (Table 1). No significant difference was found for sex, age of appearance of clinical symptoms (such as bellyache, hematochezia, anemia, etc.), and intussusception history between the two groups (p > 0.05). The patients with family history of PJS or living in rural areas have a higher incidence of cancer compared to patients without family history or living in urban areas, respectively (p < 0.05).

Table 1.

Baseline characteristics of 336 PJS patients.

	Total included (N = 336)	Patients with cancer (n = 52)	Patients without cancer (n = 284)	p value
Sex (male/female)	181/155	25/27	156/128	0.362
Age of appearance of clinical symptoms	14.5 (1–60)	15 (6–43)	14 (1–60)	0.512*
Family history (with/without)	150/186	34/18	116/168	0.001
Intussusception history (with/without)	237/99	42/10	195/89	0.078
Residence (urban/rural)	174/162	15/37	159/125	<0.001

PJS: Peutz–Jeghers syndrome.

*

Difference in age of appearance of clinical symptoms was calculated by independent sample t test.

Cancer spectrum

As shown in Table 2, among the patients diagnosed with cancer, there were 35 cases of gastrointestinal cancers (20 colorectal, 8 small intestinal, 6 gastric, and 1 pancreatic cancers) and 15 cases of extragastrointestinal cancers (4 ovarian, 4 lung, 3 cervical, 2 breast, 1 bone cancers, and 1 lymphadenoma). Two patients developed independent primary cancers (one patient with colonic and hepatic cancers and the other patient with colonic and lung cancers). The situations of sex, family history, intussusception history, and residence with regard to each specific type of cancer are shown in details in Table 2. Chi-square analysis showed a significant increase in extragastrointestinal cancers in females compared with males (p < 0.05), while no significant difference was detected for gastrointestinal versus extragastrointestinal cancers between patients with family history and those without family history, patients with intussusception history and those without intussusception history, as well as patients living in urban areas and those living in rural areas, respectively (p > 0.05).

Table 2.

The general situation of PJS patients with cancers.

Cancer	Male/female	With/without family history	With/without intussusception history	Urban/rural residence
Total	25/27	34/18	42/10	15/37
Gastrointestinal	19/16	21/14	28/7	11/24
Colorectum	9/11	14/6	16/4	6/14
Small intestine	7/1	3/5	6/2	3/5
Stomach	3/3	4/2	5/1	2/4
Pancreas	0/1	0/1	1/0	0/1
Extragastrointestinal	4/11	11/4	12/3	4/11
Ovarian	0/4	3/1	3/1	2/2
Lung	2/2	2/2	3/1	0/4
Cervix	0/3	2/1	2/1	1/2
Breast	0/2	2/0	2/0	1/1
Bone	1/0	1/0	1/0	0/1
Lymphadenoma	1/0	1/0	1/0	0/1
MPMT	2/0	2/0	2/0	0/2
p valuea	0.017	0.707	0.997	0.928

PJS: Peutz–Jeghers syndrome; MPMT: multiple primary malignant tumors.

a

Difference in gastrointestinal versus extragastrointestinal cancers was calculated by chi-square test. The two cases of MPMT were not included in the analysis.

The median age at first cancer diagnosis was 41 years (range: 21–67). Figure 1(a) showed the age distribution of the 52 PJS patients with cancers, while Figure 1(b) showed the age distribution regarding sex (male/female), family history (with/without), intussusception history (with/without), and residence (urban/rural). There was a high incidence of cancers at the age of 30–50 years old, of which the frequency was 74.5% (p < 0.001). Independent sample t test showed no significant difference in all the four comparisons (p > 0.05).

Figure 1.

Age distribution of the 52 PJS patients with cancers. (a) Age distribution of the 52 PJS patients with cancers. (b) Age distribution regarding sex, family history, intussusception history, and residence.

[Figure omitted. See PDF]

Relative cancer risks

A significantly higher RR for cancer (63.858, confidence interval (CI): 47.514–85.823) was found in the PJS patients than in general Chinese population (Table 3). There was a significantly higher risk in females (RR: 83.134, CI: 54.892–125.905) than in males (RR: 50.240, CI: 32.936–76.634) affected by any cancer (p < 0.05), while patients living in rural areas (RR: 97.268, CI: 67.400–140.371) had a significantly higher risk than those living in urban areas (RR: 31.001, CI: 18.258–52.636; p < 0.05).

Table 3.

Relative cancer risks in patients with Peutz–Jeghers syndrome.

Cancer	RR (CI)	p valuea
Total (n = 52)	63.858 (47.514–85.823)	<0.001
Gastrointestinal
Colorectum	237.918 (154.417–366.572)	<0.001
Small intestine	2600.601 (1260.315–5366.217)	<0.001b
Stomach	50.194 (22.387–112.538)	<0.001
Pancreas	41.001 (5.759–291.923)	0.024
Extragastrointestinal
Ovarian	151.538 (56.540–406.149)	<0.001
Breast	28.226 (7.030–113.329)	0.002
Cervix	69.505 (22.303–216.605)	<0.001
Lung	22.906 (8.547–61.392)	<0.001
Bone	155.467 (21.835–1106.944)	0.006
Lymphadenoma	44.684 (6.276–318.145)	0.022

RR: relative risk; CI: confidence interval.

a

Compared with data of general Chinese population.

b

Compared with data of Beijing population.

In regard of each specific site of cancers, RR for colorectal cancer was the highest (237.918, CI: 154.417–366.572) for both males (199.768, CI: 108.568–367.580) and females (289.056, CI: 156.569–533.653). Ovarian cancer was of the highest RR (151.538, CI: 56.540–406.149) among gynecological cancers. Detailed data of RR for each specific cancer were shown in Table 3.

Cumulative cancer risks

We next analyzed the cumulative cancer risks for PJS patients using Kaplan–Meier failure estimate. The cumulative cancer risk for any cancer was 4% (CI: 2–8), 21% (CI: 14–30), 47% (CI: 40–59), and 55% (CI: 44–67) at the age of 30, 40, 50, and 60 years, respectively (Figure 2(a)). For gastrointestinal cancers, the risk was 3% (CI: 2–7), 14% (CI: 9–22), 35% (CI: 25–47), and 42% (CI: 31–56) at the age of 30, 40, 50, and 60 years, respectively (Figure 2(b)). The cumulative cancer risk for colorectal cancer, the most commonly affected site of malignancy in the gastrointestinal tract, rose from 3% at the age of 30 years to 28% at the age of 60 years, as shown in Figure 2(c). For extragastrointestinal cancer, the cumulative cancer risk was 0% (CI: 0–0), 5% (CI: 2–11), 7% (CI: 3–14), and 10% (CI: 4–22) at the age of 30, 40, 50, and 60 years, respectively (Figure 2(d)).

Figure 2.

The cumulative cancer risks for PJS patients. (a) Cumulative risk for any cancer. (b) Cumulative risk for gastrointestinal cancers. (c) Cumulative risk for colorectal cancer. (d) Cumulative risk for extragastrointestinal cancers.

[Figure omitted. See PDF]

Statistically significant difference was found in the cumulative cancer risk for cancer between patients with family history and those without family history, as well as patients living in rural areas and those living in urban areas (log-rank test, p < 0.05; Figure 3(b) and (d)), while there was no statistically significant difference between males and females and patients with or without intussusception history (log-rank test, p > 0.05; Figure 3(a) and (c)).

Figure 3.

Difference in cumulative cancer risk for any cancer. (a) Sex specific. (b) Family history specific. (c) Intussusception history specific. (d) Residence specific.

[Figure omitted. See PDF]

Discussion

Numerous cohort studies as well as meta-analyses have been reported with the aim to precisely assess the risks, age of onset, and spectrum of cancers in PJS patients. However, due to the rareness of PJS and the heterogeneity of patients, there is a wide variability in estimating cancer risks.¹² RRs of 9–18 for any cancer have been reported by several different studies, with cumulative risks ranging from 37% to 93% at the age of 60–70 years.¹² Here, we found an RR for any cancer of 63.858 (CI: 47.514–85.823) with cumulative risks of 3%, 21%, 47%, and 55% at the age of 30, 40, 50, and 60 years, respectively. When cancers were analyzed regarding the specific site, colorectal cancer was of the highest risk for PJS patients of both the genders, while ovarian cancer was of the highest RR among gynecological cancers. It differs from the cohort study directed by Giardiello et al.²¹ in the United States in 2000, in which small intestinal cancer was of the highest risk for both male and female PJS patients and ovarian cancer was of the highest RR among gynecological cancers. Also, a collaborative cohort study in Italy in 2012 reported that the cancer risk was particularly increased for pancreatic and cervical cancers.²² Furthermore, the above Italian study found that there was a statistically significant difference in the cumulative cancer risk between males and females (log-rank test, χ² = 8.72, p = 0.0031), whereas in our study no statistically significant difference had been found between the two genders. We hypothesized that this kind of difference may derive from the genetic polymorphism among different races, as well as the different lifestyles among different countries. It is worth noticing that statistically significant difference was found between patients living in rural areas and those in urban areas (log-rank test, χ² = 7.34, p = 0.0067) in our study. Several factors may contribute to the differences between urban and rural people—(1) Diet: In China, the rural people like the pickled food more than urban people. In addition, the rural people eat leftovers and food that may spoil more frequently due to frugality. The above diet in rural areas may contribute to cancer risk. (2) Health and medical conditions: Generally, health and medical conditions are worse in rural area than in urban area. Besides, wealthier urban people usually pay more attention to and spend more on health. (3) Lifestyle: Urban people usually work regularly and thus could reasonably arrange time for other things, such as physical exercise, which may be beneficial for health. However, a few rural people have formal work. They usually do different work to earn money; working time was usually much longer and rather irregular. Such living may not be beneficial for their health. Of course, different ethnicities in those living in urban versus rural areas may contribute to the difference too. However, in this study, all the patients included were of Han nationality and thus, ethnicity should not be an influencing factor in the study. Nevertheless, the genetic polymorphism among different races and the different lifestyles among different countries should be known when referring data from various studies for building up cancer surveillance programs and developing therapeutic strategies. In the future, further evidences are of demand to validate these findings and, more importantly, to explore the mechanism of carcinogenesis associated with PJS.

In 1997, the PJS locus was localized on chromosome 19p13.3,²³ and in 1998, mutations in the LKB1 gene (official designation by the Human Genome Organization as STK11) at that locus were identified in PJS patients by two groups.^4,5 To date, it is still the only acknowledged gene associated with PJS, which can be found in 75%–94% of PJS patients.⁶ The existence of a second gene locus still remains debatable.⁸ Among the participants in this study, only 36 patients were analyzed for STK11 mutation. However, none of these patients has developed malignant tumor (data not shown), indicating that STK11 mutations may not lead to significant differences in the risks in Chinese population as previously reported in other countries. In fact, several large cohort studies in other countries have reported no significant differences in the risks for any cancer between patients with STK11/LKB1 mutations and those with no detectable genetic changes.^6,23,24 Nevertheless, we are now working on collecting more data for the STK11 mutation to further analyze whether there are differences between PJS patients with and without the detectable STK11 mutation in China.

Although the benefits of surveillance remain to be established, surveillance of cancer seems justified.^25,26 Several general or evidence-based guidelines have been published in recent years, as reviewed by Van Lier et al.¹² However, the rareness of PJS and wide range in reported cancer risk estimates lead to heterogeneity in surveillance protocols across countries.²⁷ Our data highlight the impact of colorectal cancer and ovarian cancer in terms of PJS-related morbidity and mortality, indicating that colorectum surveillance could be considered advisable for PJS patients. Also, surveillance for the genital tract should also be suggested to female PJS patients. Particularly, more attention needs to be given to PJS patients living in the rural areas. Moreover, besides cancers, surveillance for benign polyps should also be recommended, as complications such as intussusceptions could cause considerable morbidity and even mortality. Currently, there are already standard-of-care guidelines for PJS cancer surveillance.² We think that some current standard-of-care guidelines for PJS cancer surveillance, such as polypectomy of polyps >1 cm, may also be suitable for the Chinese population. However, indicated by the findings of our study and previously published reports, the difference also existed between Chinese PJS patients and those in other countries, such as the different high-risk cancer types. Therefore, we suggested the establishment of China’s own surveillance program for PJS patients considering these special circumstances and the work should be done as soon as possible.

Conclusion

In conclusion, in accord with previously reported studies, our 336 cases cohort study shows that PJS patients carry significant risks of cancer at various sites, with particularly elevated risks of colorectal cancer for both genders as well as high risk of cervical cancer for female PJS patients in China. Statistically significant differences are found in the cumulative cancer risk for any cancer between patients living in rural areas and those in urban areas, as well as between sporadic and familiar PJS patients. On the contrary, no significant effects of gender and intussusception history on the cumulative cancer risk are found. Hopefully, our study may help in the management of this rare disorder and establishment of related surveillance projects in China.

Compliance with ethical standards

This study was conducted with the approval of the ethics committee of Clinical College of Air Force General Hospital, Anhui Medical University. Written informed consent was obtained from all patients.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Foundation of Natural Sciences, China (No. 81470818), National Key Research and Development Program of China (No. 2016YFC0107005), Beijing Municipal Science & Technology Commission (No. D101100050010037), and Research Project of Air Force General Hospital, People’s Liberation Army (PLA) (No. kz2014020 and No. kz2015026).