Abstract
Familial adenomatous polyposis is a rare autosomal dominant inherited disease (incidence, 1/8000). More than 90% of families affected by familial adenomatous polyposis have a mutation in the tumor suppressor gene adenomatous polyposis coli (APC). Mutations in this gene are characterized by 100% penetrance, although there is a variation in phenotypic expression of the disease. According to a 2004 survey of the Italian Human Genetic Society, about 264 APC gene molecular genetic tests were performed by Italian laboratories per year. The Italian External Quality Assessment (IEQA), financially supported by the Ministry of Health and coordinated by the Istituto Superiore di Sanità, was started in 2000 to improve the quality of molecular genetic tests in Italy. In the frame of the IEQA, about 50% of public laboratories performing APC gene tests have been monitored. The number of responding public laboratories during the 5 years was 6, 7, 7, 7, and 5 from 2001 to 2006, respectively; on average, 96.3% of samples completely analyzed were correctly genotyped. Methods used by laboratories to detect mutation were direct sequencing, single-strand conformation polymorphism, protein truncation test, and denaturing high-performance liquid chromatography. Written reports were not homogeneous among laboratories, although a new form of written report was proposed to laboratories in 2004. It will be interesting to monitor the effects of the reporting model and the output of this educational action in the future.
Introduction
F
FAP is clinically characterized by early onset of multiple (hundreds to thousands) adenomatous polyps of the colon, and in its classical form, almost 100% of these polyps are considered to be cancer prone (Davidson, 2007). Rare cases of attenuated form of the syndrome with small numbers of polyps are also known (Spirio et al., 1992).
About 90% of families affected by FAP have germline mutations in the adenomatous polyposis coli (APC) gene (MIM no. 175100) located on chromosome 5q21 (Herrera et al., 1986; Bodmer et al., 1987; Groden et al., 1991; Joslyn et al., 1991; Kinzler and Vogelstein, 1991; Nishisho et al., 1991).
APC is a tumor suppressor gene; mutations at this locus are characterized by 100% penetrance, although with a variation in the phenotypic expression of the disease. Inactivation of the APC gene product constitutes the initial step in the development of colorectal cancer in FAP patients (Galiatsatos and Foulkes, 2006).
APC gene has an 8535 bp open reading frame; it consists of 15 transcribed exons encoding a 2843 amino acid protein (Fearnhead et al., 2001; Van Es et al., 2001).
The exon 15 comprises more than 75% of the coding sequence of the APC gene and it is the most affected for both germline and somatic mutations (Beroud and Soussi, 1996).
More than 935 different mutations have been found to date in the APC gene (Beroud et al., 2000) and most of them are nonsense or frameshift mutations resulting in loss of the functional protein (Beroud and Soussi, 1996; Fearnhead et al., 2001).
Hot-spot mutations are known, particularly at position 1061 as germline mutation, at 1450 as somatic mutation, and at 1309 as both germline and somatic mutations (Beroud and Soussi, 1996).
Some evidence suggests a genotype-phenotype correlation; position of the APC mutation may confer a different risk of relevant disease manifestations. The majority of mutations, correlated with classical FAP, occurs between exon 5 and the 5′ portion of exon 15, whereas those associated with attenuated FAP tend to cluster in the extreme 5′ portion of the gene and 3′ of exon 15 (Hedge et al., 2006).
FAP can be clinically diagnosed when patients develop more than 100 adenomatous polyps in their colorectum during their second and third decade of life (Nieuwenhuis and Vasen, 2007).
In a clinical setting, the aim of APC molecular analysis is the ascertainment of the genotype of the individual at risk before clinical manifestation of the diseases; further identification of a specific APC mutation can address the choice of surgical treatment. Otherwise, the genetic test offers the opportunity to confirm a suspected diagnosis (Cama et al., 1997).
There are several categories of patients appropriate for genetic testing for APC mutations.
One group consists of patients suspected of FAP based on clinical diagnosis (Davidson, 2007). Clinical findings are important because up to 25% of cases of FAP are caused by spontaneous APC gene mutations, and thus patients will therefore have no family history for the disease (Doxey et al., 2005); so far, molecular testing has been helpful in equivocal cases. A second group consists of patients with clinical evidence and family history for FAP; in this instance, identification of the mutant allele in the proband could be useful for screening other members of the family to exclude noncarriers from programs of surveillance. A third group consists of patients with a family history for FAP, but without clinical evidence of disease, and in which the mutant allele is known (Kaz and Brentnall, 2006; Davidson, 2007).
Moreover, recently published guidelines may be helpful in the appropriate management of FAP families (Vasen et al., 2008).
Genetic testing for FAP typically involves DNA sequencing, often preceded by complementary methods (prescreening test) such as protein truncation test (PTT), heteroduplex analysis (HD), single-strand conformation polymorphism (SSCP), denaturing gradient gel electophoresis, and denaturing high-performance liquid chromatography (DHPLC) (Kaz and Brentnall, 2006). Direct DNA sequencing is the most expensive but most accurate genetic test, which correctly identifies up to 95% of mutations (Kaz and Brentnall, 2006).
Once a specific genetic mutation has been found in the proband, other family members can be tested for the same mutation with virtually 100% accuracy (Grady, 2003). In families that have not undergone prior testing for FAP, genetic tests are not 100% accurate, even with direct DNA sequencing. Disease-causing mutations in the APC gene are detectable in about 85% patients with phenotypic classical FAP; moreover, some patients with negative family history or with features consistent with an autosomal recessive pattern of inheritance could carry mutations in the MYH gene (Davidson, 2007).
In this context, it is essential that the test itself gives accurate results and that the laboratory is competent to perform tests at an acceptable standard and to make sense of the information provided by the test itself (Quality and safety in genetic testing: an emerging concern; www.who.int).
External quality assessment (EQA) schemes are the main tools enabling laboratories to measure the quality of their results, to maintain confidence in molecular genetic tests, and to implement the highest standard of quality assurance (Ramsden et al., 2006; Sciacovelli et al., 2006; Guidelines for Genetic Tests, 1999).
Moreover, as APC is a large gene with a large mutational spectrum, specific competences in the molecular analysis of the APC gene are required because of reasons of time and laboratory costs (Cama et al., 1997).
The Italian EQA (IEQA) for FAP (APC gene) was started in 2001, within the IEQA in classical cytogenetics and molecular genetics (Taruscio et al., 2004); it was financially supported by the Ministry of Health and coordinated by the Istituto Superiore di Sanità (ISS).
About 264 APC gene molecular genetic tests (for FAP diagnosis) were performed by 13 Italian public laboratories in 2004 (Dallapiccola et al., 2006), and about 50% of them have been monitored within the IEQA-FAP.
The aims of the IEQA scheme are (i) to evaluate technical, analytical, and interpretative performances in APC gene genetic testing and (ii) to play an educational role, thus helping laboratories in improving their procedures. The results of five trials of IEQA-FAP scheme will be herewith illustrated.
Materials and Methods
Organization of the IEQA-FAP scheme
IEQA scheme was organized and coordinated by the ISS and funded by the Ministry of Health (Taruscio et al., 2004). Laboratories belonging to the Italian Public Health Service were enrolled on a voluntary basis and grouped into six working units (WUs), each with its own local coordinator; participation in IEQA was free.
The scheme was strictly anonymous and the identity of laboratories was only known to the scheme organizer. A steering committee, including the IEQA scheme coordinator and national experts, advised on the scientific context of the scheme and took decisions and educational actions for the development of the program (Taruscio et al., 2006).
The ISS provided an annual distribution of six validated DNA samples carrying APC gene mutations; all samples were distributed with technical information. Laboratories were asked to test samples using routine protocols and provide results of genotyping (raw data) and a full interpretative report in their normal laboratory style; data had to be sent to ISS by a given deadline (90 days after the distribution of samples). Laboratory results were evaluated by steering committee and national experts according to established criteria.
Annual workshops were organized; laboratory steering committees, assessors, and the scheme organizer met to discuss results from the IEQA (Taruscio et al., 2004).
Sample collection and validation
The DNA samples were obtained from lymphoblastoid cell lines collected by the ISS from WU coordinators and from biobank (Coriell Cell Repository, Camden, NJ). Two independent ISS WUs were responsible for DNA sample processing and validation.
Once extracted, DNA quality and quantity were checked by agar gel electrophoresis and spectrophotometer analysis; a DNase test was performed to examine the stability and integrity of DNA samples (Taruscio et al., 2004).
APC gene mutations were validated by direct sequencing. Mutations were proposed by the steering committee and were known to the ISS, but not to the recipient laboratories. Table 1 shows the list of submitted mutations.
APC, adenomatous polyposis coli.
Data evaluation
The steering committee evaluated raw data, genotyping, interpretation, and reporting of results, with particular attention to international nomenclature; all data were treated anonymously and the identity of each laboratory was unknown to the evaluating commission (Taruscio et al., 2004).
The evaluation of genotyping results took into account whether mutations had been correctly detected, not detected, or incorrectly detected. After the assessment, participating laboratories received specific comments about results, with suggestions to improve the analysis, if necessary. The scheme did not have a quantified system of assigning markers for genotyping and/or interpretations, and poor performance was not penalized.
Evaluation criteria
Raw data
The steering committee evaluated whether results were clearly and univocally interpretable; parameters for evaluation were (i) correct genotyping (the APC gene mutation was correctly detected and assigned), (ii) genotyping error (the APC gene mutation reported was not correctly identified), (iii) misinterpretative error (mutation was correctly identified, but not correctly reported), (iv) nomenclature not correct (the nomenclature used to report mutation was not in compliance with the international standard one), and (v) incomplete genotyping (samples not completely genotyped, or without conclusive diagnosis).
Written report
Clinical reports were assessed with respect to the accuracy of genotyping, appropriateness of the interpretation, and clerical accuracy; parameters evaluated are shown in Table 2. On the basis of this list of parameters, written reports were evaluated as complete or not complete.
The left column lists the parameters included in the assessment of reports. The right column shows the number of laboratories that sent an incomplete report for each parameter over the 5 years.
Parameter introduced from the fourth trial.
EQA, external quality assessment.
Results
During the 5-year survey, about 50% of Italian public laboratories performing APC gene test were monitored (Dallapiccola et al., 2006). The number of responding laboratories in annual trial and the numbers of samples analyzed are summarized in Table 3.
Genotyping result
Overall, 192 samples were sent to laboratories from ISS. On average, 96.3% (182/189) of samples were completely and correctly genotyped; errors in genotyping occurred on average in 3.7% (7/189) of samples. Three samples out of 192 (1.6%) were not completely genotyped for technical problems.
Three laboratories correctly and completely genotyped all the samples analyzed during the IEQA-FAP scheme experience; two of them participated constantly over the 5-year period, whereas the third one participated since the second trial.
Mutation c.4012C>T was not successfully detected by laboratory L2 in the fifth trial for technical failure (error in the set up of DHPLC temperature). Mutations c.3091delT, c.4592_4593insA, and c.4596_4597insT were not correctly assigned because of inaccurate reading of correct sequence; mutations c.1370C>G, c.994C>T, and c.3454C>T were correctly detected, but erroneously assigned for the shift of two nucleotides, on the reference sequence (GenBank accession no. M74088) (Table 4).
SSCP, single-strand conformation polymorphism; DHPLC, denaturing high-performance liquid chromatography; ISS, Istituto Superiore di Sanità.
Nomenclature was not in compliance with the internationally recognized one (www.hgvs.org/mutnomen/disc.html) (den Dunnen and Antonarakis, 2000) on average in 17% of samples completely analyzed.
Test methods
All participating laboratories used sequencing to detect mutations; laboratories used SSCP (53%), PTT (37%), DHPLC (22%), and HD (6%) (Table 5) as prescreening techniques.
Values refer to number of laboratories that used specific techniques on the total number of laboratories for each year.
PTT, protein truncation test.
The use of DHPLC increased from 2001 to 2006, whereas the use of other techniques, such as PTT, SSCP, and HD, considerably decreased during the five trials of IEQA-FAP scheme (Table 5).
The testing strategies adopted by laboratories to perform their analyses are shown in Table 6. Few laboratories used only sequencing to directly detect mutations, whereas most of them preferentially used sequencing to analyze only the region recognized by prescreening techniques.
A, only sequencing method to detect mutations; B, sequencing after a prescreening analysis by other techniques; B1, sequencing after a prescreening analysis performed by one technique; B2, sequencing after a prescreening analysis performed by two techniques; B3, sequencing after a prescreening analysis performed by more than two techniques. Prescreening techniques were PTT, SSCP, heteroduplex analysis, and DHPLC.
Reporting results
Reports sent by laboratories were not homogeneous, and in about 50% of them information was lacking (Table 7).
ISS, steering committee, and national experts elaborated a new written report model to standardize reporting information. This new model was proposed and adopted by participating laboratories in 2004 (Table 2); more complete information was requested from laboratories because new parameters were added to the previous model. The most frequent inaccuracies in reports were the lack of indication for genetic counseling and the absence of interpretation of the result, on average in 21.8% and 12.5% of written reports, respectively. Lack of analytical sensitivity and specificity of the procedures and mutation detection rate were observed on average in 41% of written reports during 2004 and 2006, respectively (from 2001 to 2003, these parameters were not checked) (Table 2).
Discussion
Genetic research has made a substantial impact on health improvement, disease diagnosis, and prevention of diseases; in the last few years, molecular genetic techniques have entered into clinical practice.
Unlike most other clinical laboratory tests, the result of genetic testing for a given disorder will not change over time and it is important to determine the clinical management of the patient and his family; the clinical consequence of an error in the molecular analysis for genetic diseases may be serious (Ramsden et al., 2006; Mackie et al., 2007).
Laboratories who offer genetic tests are asked to implement the highest standards of quality assurance and in the same time to keep in confidence with new technologies (Cama et al., 1997; Ramsden et al., 2006).
EQA is recognized as an essential component of quality assurance by governmental and nongovernmental organizations, and they were developed at both a national and international level (Dequeker et al., 2001; Ibarreta et al., 2004; Mueller et al., 2004; Taruscio et al., 2004; Hertzberg et al., 2006; Ramsden et al., 2006; Salvatore et al., 2007; Falbo et al., 2008; Floridia et al., 2008; Tosto et al., 2009).
Medical laboratories have a long tradition in the organization of EQA (Libeer, 2001). In 1947, results from the first clinical chemistry survey were published by Belk and Sunderman (1947); survey in clinical molecular genetics started in Europe at the beginning of 1990s (Ramsden et al., 2006). These activities not only are about laboratory performance evaluation, but also play an important educational role; schemes are generally designed to evaluate the performance of the method and to help laboratories in improving their performance (Libeer, 2001).
Recommendations for the molecular diagnosis of genetic disorders, such as FAP, highlight the need for competent institutional authorities to organize appropriate and periodic quality control (Cama et al., 1997).
The lack of mutation detection in at-risk individuals has consequences on their clinical management, because they could interrupt, in agreement with clinicians, annual check-ups.
From 2001, EQA schemes for molecular FAP diagnosis (APC gene) are available at a national and international level (Taruscio et al., 2004; Ramsden et al., 2006).
The UKNEQAS survey for APC started in 1991, and no errors were registered (Ramsden et al., 2006).
The IEQA scheme for FAP diagnosis (APC gene) started in 2001 and represents a 5-year experience.
In this respect, 189 samples have been completely analyzed; 96.3% of samples were correctly genotyped and no errors occurred in 2001 and 2004. About 3.7% of samples were not correctly genotyped by participating laboratories. Several studies indicate that genotyping errors occur in other EQA schemes with different design and for different genetic diseases (Dequeker et al., 2001; Mueller et al., 2004; Salvatore et al., 2007; Falbo et al., 2008; Tosto et al., 2009).
In general, quality of raw data appears to be high, and only on a few occasions it was not optimal for a correct interpretation of results.
Only one technical error was observed in the detection of c.4012C>T mutation, and it was due to incorrect setup of DHPLC screening method used. The laboratory (L2) again analyzed the sample after the evaluation of results and then it correctly detected the mutation.
The majority of errors (seven out of eight) recorded in the IEQA-FAP scheme was due to misinterpretation of correct results; we classified these errors into two groups: (i) mutations present on the raw data but not correctly reported because their identification was difficult for raw data of less than optimal quality (c.3091delT, c.4592_4593insA, and c.4596_4597insT); and (ii) mutations identified on the raw data but not correctly described because laboratories did not accurately align their results on the reference sequence and two nucleotide shifts were reported (c.994C>T, c.1370C>T, c.3091delT, and c.3454C>T) correctly.
Technical failure could be minimized by detecting mutations with more than one technique and by confirming results with DNA sequencing on both strands (Mueller et al., 2004).
During the five trials, laboratories used different techniques to perform analyses of the APC gene. Almost all of them (74.2%) used sequencing on the mutation region identified with prescreening tests, whereas the remaining 25.8% used this technique to directly test mutations. It is important to note that Italian assessors encouraged the use of prescreening techniques to minimize expensive and time-consuming approaches.
Regarding prescreening techniques, an increase in the use of DHPLC among laboratories was observed; consequently, the number of laboratories using SSCP and PTT decreased, and in the last 2 years, HD was not used.
Laboratories were asked to send to ISS their commonly used reports to assess and to get better quality. As previously observed in other IEQA schemes, reports were not homogeneous, and incompleteness was often detected (Salvatore et al., 2007; Falbo et al., 2008; Floridia et al., 2008; Tosto et al., 2009).
A new model for reporting was elaborated in 2004 from ISS, steering committee, and national experts. After the introduction of the model the number of incomplete reports increased because the criteria of assessment were more stringent, and some important information was frequently missed by laboratories, particularly the indication for genetic counseling, the interpretation of the results, analytical sensitivity and specificity of the procedures, and mutation detection rate of the test. It will be interesting to monitor the output of this educational action over the next few years.
As shown by other EQA programs, constant participation to the schemes has been linked to improved laboratory performance in genetic testing, even if an intrinsic analytical error of the tests will be present among participating laboratories.
Finally, a great contribution to the educational process of testing comes from the annual workshops organized within the frame of IEQA and during which the steering committee and participating laboratories reviewed results from the scheme.
Footnotes
Acknowledgments
This work was funded by “Progetto per la standardizzazione e l'assicurazione di qualità dei test genetici” (2000-2002) and “Test genetici: dalla ricerca alla clinica” (2003-2006), Italian Ministry of Health. The authors are grateful to Dr. Manuela Marra and Dr. Fiorentino Capozzoli for critical reading and comments on this manuscript. The authors thank the laboratories that participated in the IEQA, for their contribution of data to the scheme.
Disclosure Statement
No competing financial interests exist.