Dental Screening Including Panoramic Radiograph for Gorlin-Goltz Syndrome in Patients With Multiple Basal Cell Carcinomas

Abstract

Purpose

To answer the following clinical research question: “Among patients with multiple basal cell carcinomas (mBCCs), can panoramic radiograph (PaR) facilitate the diagnosis of Gorlin-Goltz syndrome (GGS)?”

Methods

This retrospective study enrolled mBCCs subjects who presented to a German tertiary care center between 1 January 2015 and 31 December 2021. The primary predictor was presence of syndromic mBCCs, and the main outcomes were jaw cysts and odontogenic keratocysts (OKCs). Descriptive, bi- and multivariate statistics, diagnostic test evaluation, and number needed to screen (NNS) were computed at α = 95%.

Results

The sample comprised 527 mBCCs patients (36.1% females; 6.8% GGS; 5.5% OKCs; mean age, 74.5 ± 15.8 years [range, 15-102]). There was a significant association between syndromic mBCCs and jaw cysts (P < .0001; NNS = 2 [95% CI, CI, 1.1 to 1.4]). In the adjusted logistic model, PaR identified GGS via radiographic diagnosis of jaw cysts in case of 1) age ≤ 35 years, 2) ≥ 5 BCCs, and 3) ≥ 1 high-risk BCCs. Nearly every jaw cyst identified by PaR was OKCs (P = .01; 95% CI, 3.1 to 3,101.4; NNS = 1.3 [95% CI, .9 to 2]). The post hoc power was 100%.

Conclusions

Dental screening with the use of PaR for mBCCs patients, especially those aged ≤35 years, or with ≥5 BCCs, or ≥1 high-risk BCCs, may be helpful in detection and identification of GGS through recognition of OKCs.

Introduction

Although Gorlin-Goltz syndrome (GGS, or nevoid basal cell carcinoma syndrome; OMIM#109400) is not as common as cleft lip/palate (prevalence: 1:31,000 to 1:256,000 vs. 1:500), life expectancy with GGS is significantly lower than that of the general population (median: 73.4 [95% confidence interval, CI, 64.0 to 82.8] vs. 80 [95% CI, 79.8 to 80.2] years; 10-year survival, 93.3 ± 6.4%). ^1,2 Moreover, it has neoplastic predisposition, e.g. basal cell carcinomas (BCCs) and extracutaneous diseases including odontogenic keratocysts (OKCs), dyskeratotic palmar and plantar pitting, and skeletal and developmental anomalies. ^{1 -3}

Much uncertainty still exists about the diagnosis of GGS, i.e. no single manifestation supporting the diagnosis. The recently updated 5th Edition of the World Health Organization (WHO)’s classification of familial head and neck tumor syndromes confirms the unchanged diagnostic criteria used over a decade (Table 1). ³ Although the presence of multiple OKCs (mOKCs) is one major criterion, there remains a paucity of evidence on which and when mBCCs patients should undergo dental screening including the use of panoramic radiograph (PaR) for detecting GGS-related mOKCs.

Table 1.

Diagnostic Criteria for Nevoid Basal Cell Carcinoma Syndrome (GGS) After the 2022 Updated 5th Edition of the World Health Organization Classification of Familial Head and Neck Tumor Syndromes, ³ and Modifications by Krakowski et al. ⁶

Major criteria

Minor criteria

–Basal cell carcinoma before age 20, or excessive numbers of basal cell carcinomas out of proportion to prior sun exposure and skin type –Odontogenic keratocyst before age 20 –3 or more palmar or plantar pitting –Lamellar calcification of the falx cerebri –Medulloblastoma, typically desmoplastic –First degree relative with nevoid basal cell carcinoma syndrome

–Bifid, fused, or markedly splayed ribs –Other specific skeletal malformations and radiologic changes (e.g., hemivertebrae, fusion, or elongation of the vertebral bodies such as kyphoscoliosis, sprengel deformity, marked pectus deformity, short fourth metacarpals, postaxial polydactyly) –Macrocephaly –Cleft lip or palate (or any other congenital malformation of the craniofacial region, e.g. frontal bossing, coarse face) –Ovarian or cardiac fibroma –Lymphomesenteric cysts –Ocular abnormalities (i.e., strabismus, moderate or severe hypertelorism, congenital cataracts, glaucoma, coloboma)

Note: Requirements for diagnosis included (1) two major diagnostic criteria and one minor diagnostic criterion, (2) One major and three minor diagnostic criteria, or (3) identification of a heterozygous germline PTCH1 or SUFU pathogenic variant on molecular genetic testing.

The present investigators asked the research question: “When a patient presents with mBCCs, can PaR be used to identify syndromic mBCCs with mOKCs?” In this paper, the term “syndromic” was referred to GGS-associated. We hypothesized that PaR could be used as a proxy variable for GGS, and that there exists a set of predictors of GGS. The specific aims of the study were 1) to design and implement a retrospective cohort study on a sample of syndromic and sporadic mBCCs patients with/without mOKCs, and 2) to identify clinical usefulness of PaR on the GGS diagnosis. The results of this study will provide the 2011 Oxford Centre for Evidence-Based Medicine (OCEBM)’s level of evidence “3 (screening)”, and recommendation grade “B”.

Patients and Methods

Study Design and Sample Description

The investigators organized a retrospective monocentric cohort study of mBCCs cases presented to the Craniomaxillofacial Surgery (CMFS) and Plastic Surgery units in a German tertiary medical center between 1 January 2015 and 31 December 2021. Subjects were selected from the hospital’s pathological registry and were eligible for study inclusion if there was a final histological diagnosis of mBCCs with/without mOKCs. Subjects with unavailable or incomplete medical records, or inadequate PaR information were excluded.

The institutional review board approved the project, and all patients gave consent to the use of their anonymous data. There was neither human contact nor tissue manipulation during this study, i.e. chart reviewing only. The Helsinki Declaration’s ethical guidelines and the STROBE statement were adhered throughout the study.

Study Variables

The primary predictor variable was syndromic mBCCs with the diagnosis of GGS being confirmed according to the abovementioned WHO’s classification (Table 1). ³ It was a binary variable (yes vs. no).

The main outcomes were presence of jaw cysts on PaR, and accuracy of radiologic OKC diagnosis, which were recorded in binary (presence vs. absence of jaw cysts; and correct [i.e., histologically confirmed] vs. incorrect OKC diagnosis by PaR [i.e., histological diagnosis of other cysts]). Syndromic OKC(s) present as multiple (≥ 2) lesions, especially in the posterior mandibular area, in GGS patients (and in this study, with mBCCs). They can be either unilocular (≥ 80%, i.e. no septum within the lesion) or multilocular, and in many cases, grow mesiodistally without the buccolingual expansion. ^4,5 Pictorial details about radiological features of OKCs (n = 120) were previously described by Pitak-Arnnop et al. ⁴

Additional data were collected regarding demographics: 1) age (continuous data, and adjusted into binary: ≤35 vs. >35; Rationale: syndromic mBCCs appear most commonly between puberty and 35 years of age, unlike the older age that is seen with sporadic mBCCs ⁶ ), and 2) gender (female vs. male), and clinical-pathological features: 1) number (2-4 vs. ≥5; Rationale: the average of total lifetime BCCs is ~5, and median of ~3 [ranges, 0-85 ⁷ ; 2-50 ⁸ ), and 2) ≥1 high-risk BCC (yes vs. no). We followed the risk stratification of BCCs based on the German S2k (consensus-based) practice guideline (Table 2). ⁹

Table 2.

“High-Risk” (For Recurrence and Metastasis) BCCs Based on the Current German S2k (Consensus-Based) Practice Guideline (Valid From 1 June 2018 to 31 May 2023). ⁹

“High-risk” BCC

≥ 2 mm in thickness horizontal diameter > 6 mm for BCCs on the central face (eyelids, eyebrows, periorbital area, nose, upper lip, mandibular angles, pre- and postauricular region, ear, and temple), urogenital skin, hands, and feet horizontal diameter >10 mm for BCCs on the cheeks, forehead, chin, lower lip, scalp, neck, and pretibial skin horizontal diameter > 20 mm for BCCs of trunk and extremities ill-defined margins histologic sclerodermiform, infiltrative, metatypical, or micronodular subtypes invasion into the lower dermis or subcutis layers of the skin invasion into the tiny nerves in the skin (perineural invasion) a combination of these features

Data Management and Statistical Analysis

Data were collected from the hospital’s pathological registry by an external (non-surgeon) author (L.-K.W.), and transferred as blinded data to a Microsoft Excel 2007 worksheet (Microsoft Inc., WA, USA). Descriptive statistics were computed for all variables, and bivariate analyses were used to measure the associations between the study variables at α = 95%. Characteristics associated with the outcomes of interest in bivariate analysis (P < .1) were included in a multiple logistic regression model to investigate whether such characteristics predicted GGS. All significances of differences were assessed with the help of MedCalc^® (MedCalc Software Ltd., Ostend, Belgium).

The number needed to screen (NNS) is defined as the number of people that need to be screened for a given duration to prevent one death or adverse event. It equals 1 divided by absolute risk reduction. ¹⁰ Based on the “rule of thumb”, single-digit NNS confirms the speculation that the screening is advantageous. ¹¹

The post hoc power based on the main therapeutic outcome, i.e. the presence of jaw cysts on PaR, were computed using G Power 3 for Windows (HHU Düsseldorf, Düsseldorf, Germany) with an effect size of .5, an α error probability of .05, and a sample size of 527.

Results

During the study interval, 527 mBCCs patients (36.1% females; 6.8% GGS; 5.5% syndromic OKCs; mean age, 74.5 ± 15.8 years [range, 15-102]) were identified; no eligible subject was excluded. Supplemental Table 1 in the supplement file summarizes the descriptive and binary statistics and NSS calculations of the study variables. In brief, syndromic mBCCs were significantly associated with age ≤ 35 years (P < .0001; odds ratio [OR], 438.4; 95% CI [CI], 55.4 to 3,468.3), female gender (P = .03; OR, 2.1; 95% CI, 1.1 to 4.1), ≥ 5 BCCs (P < .0001; OR 66.1; 95% CI, 22.5 to 194.3), ≥ 1 high-risk BCCs (P < .0001; OR, 10.7; 95% CI, 4.8 to 24.1), and ≥ 1 jaw cysts (P < .0001; relative risk [RR], 20.1, 95% CI, 12.9 to 31.3; NNS = 2 [95% CI, 1.1 to 1.4]). Nearly every jaw cyst identified by PaR was histopathologically confirmed as an OKC (P = .01; OR, 98.3; 95% CI, 3.1 to 3,101.4; NNS = 1.3 [95% CI, .9 to 2]).

Using PaR findings of OKCs to diagnose syndromic mBCCs yielded the accuracy of 68.5% (95% CI, 49.73% to 83.69%), and posterior probability of positive test (W_p) was 97% (95% CI, 88% to 99%) and of negative test (W_n) was 0% (95% CI, 0% to 82%) (Supplemental Table 2 in the supplement file), i.e., every patient with a positive PaR result had an OKC, and every negative PaR patient was free from the OKC.

The multiple logistic regression model for the presence of GGS was summarized in Supplemental Table 3 (in the supplement file). After controlling for the effects of syndromic mBCCs, age ≤ 35 years, ≥5 BCCs and ≥1 high-risk lesion were statistically associated with the GGS diagnosis (adjusted r ² = .47; degree of freedom, 4; P = .000). For increasing age, the figures were OR 1,170.4, and 95% CI, 63.1 to 21,700.9 (P = .000). Patients with ≥ 5 BCCs had a 223.2-fold increase risk of having GGS over those with 2 to 4 BCCs (OR, 223.2; 95% CI, 25.2 to 1,975.4; P = .000), and owning ≥ 1 high-risk lesions was associated with a 48.1-fold (OR, 48.1; 95% CI, 5.5 to 418.5; P = .0006) increased risk of having GGS.

Considering the β-coefficients (Supplemental Table 3 in the supplement file), every increased unit of the three significant predictors: number of high-risk lesions, number of BCCs, and year of age, will increase ~4, ~5, and ~7 predicted syndromic mBCCs patients, respectively (β-coefficient: number of high-risk lesions, 3.9; number of BCCs, 5.4; age, 7.1).

The post hoc power was 100%, confirming 100% chance of our results with a real effect.

Discussion

Recently, there has been a considerable increase in literature pertaining to GGS. Most of published data are, however, limited to case reports or series, and narrative reviews including consensus-based guidelines. What is less clear is the interdisciplinary approach to GGS patient care (including its benefit-risk profile), for example, primary preventive measures for less or invisible organs to identify possible abnormalities such as calcification and tumors of the brain, and jaw cysts. mBCCs patients may suffer from a variety of syndromes (Supplemental Table 4 in the supplement file). Identifying OKCs will not only aid in confirming GGS diagnosis (as a major diagnostic criteria [Table 1]), but promote a prompt curative action before the cyst becomes much larger, difficult to treat with increase potential for recurrence. Syndromic OKCs tend to recur much more often than sporadic cysts. Moreover, left undiagnosed or untreated, OKCs may result in substantial bone loss and subsequent pathologic fractures. ^4,12 Notwithstanding its importance, dental screening for GGS has received scant attention in the analytic literature.

The aims in the present study were 1) to assess the clinical usefulness of PaR in GGS diagnosis among mBCCs patients, 2) to use a defined set of variables to develop a multiple logistic regression model to predict GGS in this patient population, and 3) to validate this model. We hypothesized that PaR ameliorated GGS diagnosis in patients with mBCCs and jaw cysts, and that there may be a set of predictors of GGS. Our main findings highlighted the significant association between syndromic mBCCs and jaw cysts (P < .0001; NNS = 2 [95% CI, 1.1 to 1.4]). mBCCs patients with at least one jaw cyst had 20-fold increased odds (OR, 20.1; 95% CI 12.9 to 31.3) of GGS, when compared to those with no jaw cyst. Using multivariable analysis, three factors associated with syndromic mBCCs patients with a jaw cyst were 1) age ≤ 35 years (OR, 1,170.4; 95% CI, 63.1 to 21,700.9; P = .000), 2) ≥ 5 BCCs (OR, 223.2; 95% CI, 25.2 to 1,975.4; P = .000), and 3) ≥ 1 high-risk BCCs (OR, 48.1; 95% CI, 5.5 to 418.5; P = .0006).

Because of its rarity, much of the research up to now has been descriptive in nature. Two of the present study’s authors (J.-P.M., P.P.) reported a 10-year series of odontogenic cysts (n = 695) treated at Pitié-Salpêtrière University Hospital in Paris, France, and found 120 sporadic and 10 syndromic OKCs. Two (or 20%) of syndromic OKCs developed before other manifestations. ^4,13,14 In other words, ~1 new syndromic OKC per year was treated, and sporadic OKCs were 12-fold more often than syndromic cysts (P = .058; 95% CI, −.03 to 1.7). In comparison with the present cohort, there were 29 OKCs in 36 syndromic mBCCs patients, suggesting ~4 new syndromic OKCs per year. A possible explanation is that the Parisian region has more CMFS units (i.e., in 5 hospitals) than most German cities, which generally have only one hospital-based CMFS department (except in some larger cities, e.g. Berlin, Munich, Stuttgart, and Hannover). ^{15 -18} Many German federal states have only one hospital-based CMFS center in their capital, e.g. Dresden, Magdeburg, Wiesbaden, Mainz, Saarbrücken, and Kiel. ^{15 -18} The hospital volume of German CMFS units could, therefore, be denser in this regard. However, this doctrine is not always true. For example, an 11-year cohort from Regensburg showed 36 sporadic OKCs (or ~3 new cases per year) only, and syndromic OKC was not reported. ^19,20

In this study, the use of PaR to identify a jaw cyst is a reliable diagnostic method for screening GGS in patients with mBCCs (accuracy, 68.5% [95% CI, 49.73% to 83.69%]; W_p, 97% [95% CI, 88% to 99%]; W_n, 0% [95% CI, 0% to 82%]). Given the high sensitivity (1.0), no jaw cyst on PaR strongly excludes the possibility of a GGS diagnosis. The moderate specificity (.67) suggests that an identified jaw cyst fairly indicates a GGS diagnosis. Hence, the GGS diagnosis may primarily include jaw cyst consideration (i.e., presented at young age, multiple jaw cysts and/or recurrences ^{4,12 -14} ) without delay due to histological diagnosis. It is, therefore, far easier for non-dental health providers such as dermatologists, neurologists, neurosurgeons, internists, or family physicians, to order a radiograph such as a skull series for jaw bone evaluation in case of the absence of an in situ dentist or CMFS surgeon, regardless of the patient’s dental compliant or symptom.

The concurrence between syndromic OKCs and BCCs is linked to mutations in tumor suppressor genes PTCH1 and PTCH2, leading to continuous activation of the hedgehog pathway, ⁶ which regulates patterning and development in the embryo and adult. ²¹ A recent Stanford study observed jaw cysts in 180 of 248 (or 74.1%) GGS patients. ⁷ Our cohort included syndromic OKCs in 29/36 (or 80.6%) syndromic mBCCs patients (compared to the sporadic mBCCs group: P < .0001; RR, 784.5; 95% CI, 48.9 to 12,585.6; NNS 1.3 [95% CI, 1.2 to 1.3]). Syndromic mBCCs were also significantly associated with jaw cysts (P < .0001; NNS = 2 [95% CI, 1.1 to 1.4]), which may cause other oral abnormalities such as tooth impaction or agenesis, or dentofacial defomities. ^22,23 Based on the computed NSS, PaR screening of 10 in every 13-20 mBCCs patients would identify one GGS patient.

The 2022 WHO’s classification of familial head and neck tumors adopted the 2020 recommendations of the International Society of Pediatric Oncology. PTCH1-associated GGS type 1 patients necessitate oral examination at age 2 with annual PaR from age 8 onwards, and every 3 years after age 30, while SUFU mutation (GGS type 2) carriers have no OKCs, and require no dental screening. ^3,24 In contrast, the US National Cancer Institute (NCI) launched three recommendations on this matter, regardless of the genetic phenotype: 1) starting after the first OKC diagnosis, then every 6 months until age 21 years or until no cysts for two years, 2) beginning at age 8, then every 12-18 months if no cyst is found, and 3) waiting until symptomatic to begin PaR to limit the radiation exposure. Unfortunately, the US NCI did not conclude its most preferred screening protocol. ²⁵ To reduce radiation risk, the Dutch guideline posits biennial dental including PaR examinations in GGS-confirmed or -suspected patients from age 8 to 22. After age 22, PaR is needed only in case of pain and/or an unexplained position change of the teeth. ²⁶

The WHO’s diagnostic criteria recommend that GGS be considered, if the onset of mBCCs and/or OKCs is before age 20. ³ We tested this cutoff and found it powerful (i.e., 3 syndromic vs. 0 sporadic cases; P = .002; OR, 102.7; 95% CI, 5.2 to 2,029.7). It is worth mentioning that the genetic investigation is not always done unless GGS is suspected, e.g. in a familial history (~45% to 50% of GGS patients, ^1,7,11 and only 4/36 [or 11.1%] cases in our study, compared to the prevalence of syndromic mBCCs in GGS patients which ranges from ~50% in the Japanese to 70% to 90% in Caucasians and Koreans ^7,21,27 ). In the Stanford study, the average total lifetime BCC count was 215 ± 299 (median, 130; range, 0 to 2,200), and ~25% of syndromic BCCs were advanced and/or metastatic. ⁷ The combination of these findings supports the idea that GGS may present as mBCCs before age 20 rather than with a familial history, i.e. de novo mutations.

The age of 15 was our youngest GGS onset, while 24 of 36 (66.7%) cases developed a jaw cyst after age 30. Hence, PaR screening from the age of 8 is unlikely to be beneficial and cost-effective. Additionally, albeit low (~.01 mSv, compared to conventional computer tomography [CT], ~2.1 mSv, and cone-beam CT, ~.05 mSv), radiation exposure from PaR might predispose the patients to new BCCs. ^{1,3,5 -7,9,26} Our results could well close this clinically relevant gap. If an mBCCs patient aged ≤ 35 years, or with ≥ 5 BCCs, or ≥ 1 high-risk lesions enters the medical consultation, she or he ought to be dentally screened as early as age 15 or the first onset of mBCCs or jaw cysts, irrespective of testing of germline mutations and/or presence of other symptoms.

We are aware of many study limitations. Despite data collection by an external author (L.-K.W.), a potential bias of this study is the retrospective single-institution study design. Missed or incorrect documentation cannot be excluded. Further weakness is the difficulty in some explanations. For example, the youngest case in the present cohort was 15, while previous studies reported children with GGS at age 2-9. ¹² This contradictory result may be due to the fact that young head and neck syndrome children in Germany are often referred from pediatricians or dentists to cleft-craniofacial centers such as the CMFS department of University Hospital Frankfurt, Cologne, or Leipzig where holistic, interdisciplinary patient care have been well-established. Moreover, jaw cysts may be treated in private clinics, and not included in this study. The ratio of syndromic to sporadic mBCCs in this investigation is significantly below those observed by a US multicentric research group (hoc impleatis, 6.8% [36/527] vs. US, 10.9% [55/503]; P = .027; OR, .6; 95% CI, .38 to .93). ⁸ Patient referral bias is beyond our study scope and requires further investigations. It is also unfortunate that the PaR was not performed in all patients, as it is unnecessary for patients with sporadic mBCCs who may therefore escape from an OKC diagnosis at onset of mBCCs. This is also possible that some mBCCs patients develop OKCs in later life, and may later be diagnosed as GGS. Lastly, the molecular profiles of BCCs and OKCs, the genetic profiles of GGS, the cost-benefit analysis, and the effect of sunburns and radiation exposure were not known in this study. Differential diagnosis for mBCCs occurring at an early age is presented in Supplemental Table 4 in the supplemental file. ^1,3,5,6,28

Conclusions

Our findings demonstrate a large clinical benefit of dental examination in patients with mBCCs as early as age 15 or the first onset, especially in those with: 1) age ≤ 35 years, 2) ≥ 5 BCCs, and 3) ≥ 1 high-risk BCC. ~80% of GGS patients in this cohort have at least one OKC, and every 13-20 subjects with mBCCs would benefit from PaR screening to identify 10 GGS. The cost-effective-risk analysis of dental screening including PaR is an intriguing issue, which could be usefully explored in further research.

ORCID iD

Poramate Pitak-Arnnop https://orcid.org/0000-0002-7427-3461