Family history of cancer: It’s in the details!

Drawing and interpreting the patient’s medical pedigree is a crucial step. Most familial cancer syndromes have an autosomal dominant inheritance, and a family history of a particular type of cancer in successive generations is suggestive of such a syndrome, especially if the ages at diagnosis are relatively young. The suggestion of a familial cancer syndrome is stronger for familial clustering of rare cancer types like sarcomas than for clustering of common types like breast cancer.

The occurrence of different types of cancer within the same family may strongly suggest hereditary cancer if these combinations are known to be associated with a syndrome (3), e.g., breast and ovarian cancer and BRCA1 or 2 mutations; colorectal and endometrial cancer in Lynch syndrome; breast cancer and sarcoma in Li-Fraumeni syndrome; and diffuse type gastric cancer and lobular breast cancer in hereditary diffuse gastric cancer. Diagnosis of a cancer that predominantly affects one gender in a patient of the opposite gender — with breast cancer in males as the best known example — is another clue. Cancer diagnosis later in life never fully excludes a familial cancer syndrome as the range of ages at diagnosis are notoriously wide for most of these syndromes.

A positive cancer family history in itself is usually not enough to diagnose a cancer syndrome with certainty, especially when common type cancers are involved. Chance alone may cause the clustering of cancer (especially in large families), and an appreciation of the size of the family, the number of unaffected relatives, and their place in the pedigree is important. Shared environmental risk factors may contribute heavily to the familial occurrence of cancer, e.g., a familial habit of heavy smoking to the clustering of lung cancer, and they should be ascertained.

A negative family history of cancer does not exclude a familial cancer syndrome for several reasons. Virtually none of the germline mutations involved in hereditary cancer have a 100% penetrance. Families may be small or non-informative for other reasons including lack of knowledge of the family history, the death of parents and other relatives at a relatively young age from non-cancer causes, or the small number of females relatives of a proband with breast or ovarian cancer. Also, cancer patients may have a new, de novo, mutation and therefore be the first in their family with a syndrome. This phenomenon has been reported for many genetic disorders, most commonly for familial adenomatous polyposis and multiple endocrine neoplasia type 2B. De novo mutations are uncommon in patients with Lynch syndrome and BRCA1/2 hereditary breast-ovarian cancer. Apparent single cases in a family can also be caused by a different mode of inheritance, such as autosomal recessive, e.g., in MUTYH-associated polyposis, X-linked, e.g., in X-linked lymphoproliferative syndrome, or genomic imprinting, e.g., in SDHD gene-associated familial paraganglioma. In these single cases in a family, a relatively young age at diagnosis should still raise suspicion of a familial cancer syndrome.

A family history of cancer is not necessarily accurate (4-6) and should therefore be verified if possible before making a diagnosis. Clinical as well as pathological data are important. Pathology reports can supply important details on the classification of tumors, e.g., renal oncocytoma rather than clear cell renal cancer, which would be highly relevant to the genetic differential diagnosis (7). In addition, pathology reports may contain other clues, for example, basal-like phenotype and triple-negativity of breast cancer associated with BRCA1 mutations (8) or the tumor-infiltrating lymphocytes in colorectal cancer associated with Lynch syndrome and MUTYH-associated polyposis (9).

Rare neoplasms

Some rare cancer types have a much higher than average risk of being a manifestation of a familial cancer syndrome and clinicians should be aware of these types. Some well known examples are shown in Table 1.

Table

Table 1. Rare tumor types and their associated cancer syndrome(s). WAGR = Wilms’ tumor, aniridia, genital-urinary anomalies and retardation.

Adding it all up

The sum of all observed features in a cancer patient and his or her family leads to the next step. Depending on the results, this can range from reassuring the patient that the cancer is unlikely to be genetic, to DNA testing or other lab analysis, or to a final diagnosis on the basis of a set of clinical and pedigree criteria alone. Discussing these next steps and final diagnoses is beyond the scope of this chapter. It is important to realize that many (if not most) families with clustering of common type cancers like breast or colorectal cancer will in the end not be diagnosed with a familial cancer syndrome but can still be at significantly increased cancer risk and may be eligible for enrollment into a cancer surveillance program.

Introduction

The Familial Cancer Database (FaCD) has primarily been developed as a tool to support the genetic differential diagnosis of cancer (15, 16). In 2008, a new online version of FaCD was published and updates have subsequently been provided on a regular basis. Use of FaCD is free for health professionals after registration. FaCD is not an expert system, but rather a tool for experts and is therefore no substitute for consulting an specialist in genetic diseases.

The software tries to match tumor and non-tumor features observed in patients and their families with any of the nearly 500 disorders presently contained in its database and provides a synopsis for each of these disorders. This synopsis focuses on phenotype and cancer risk estimates. The following classes of disorders have been included:

• Hereditary disorders certainly or possibly associated with increased cancer risk
• Reported familial clustering of particular types of cancer
• Known patterns of multiple primary tumors
• Common disorders and external risk factors associated with increased cancer risk (e.g. diabetes, alcohol). These disorders have been added to the database because they can occur in familial clusters and can be associated with their own particular cancer risks

FaCD is intended to be a comprehensive resource and includes all reported syndrome tumor associations, even if not confirmed by molecular evidence. When searching the database, the user is able to include or exclude uncertain associations from the search results. We also wanted the system to allow the user to input not only very specific search terms, e.g. ‘malignant fibrous histiocytoma’, but also broad search terms like ‘renal tumor’ (whatever type) or ‘gastrointestinal tumor’ because the user might want to check the database at a time when a reported family history still lacks details. FaCD does not contain all possible search terms but only the tumor and non-tumor features that have actually been reported in association with the syndromes included in the database.

Searching FaCD for matching disorders

A typical search would go as follows:

First, the user selects symptoms from the dropdown lists in the Symptoms-by-System menu (not shown) or through the Find symptoms box (Figure 1).

Figure

Figure 1. Search results of the Familial Cancer Database (www.facd.info) using the term “colo”.

Next, the user adds symptoms of his or her choice to the Selected Symptoms box (Figure 2). Bearing in mind the typical tell-tale signs of familial cancer syndromes, most users will select the most suspect features from all those that have been observed in the family under investigation. However, FaCD offers the option of including a long list of observed features and allows the user to include or exclude a selection of them in different searches by checking or unchecking the Ignore boxes in the search profile.

Figure

Figure 2. The Selected Symptoms box used to refine queries of the Familial Cancer Database (www.facd.info)

Next, the user has FaCD run the search profile through the database specifying:

• how many of the symptoms should be present in a disorder to match the search criterion (e.g. “all” or “2 or more”)
• if only disorders that are associated with the selected symptoms should be searched for (narrow search) or also those that are possibly/probably associated (wide search).

Running the search profile will return a list of matching disorders (Figure 3).

Figure

Figure 3. Search results from the Familial Cancer Database (www.facd.info).

The user then selects disorders of interest from this list and views them in detail, with an option to print any of these detailed syndrome descriptions. The user can also return to the search profile box, make changes to the search parameters and re-run the profile.