Clinicopathologic Characteristics of Young Patients with Oral Squamous Cell Carcinoma

Abstract

Oral squamous cell carcinoma (OSCC) occasionally occurs in young patients and is likely to be distinct from OSCC in older patients. In this retrospective study, we described the clinicopathologic features and outcome of 150 OSCCs that were diagnosed in patients 40-year-old or younger. Most patients (63%) were non-smokers. The most common site of the primary tumor was oral tongue (n = 131, 87%), followed by gingiva (n = 9), buccal mucosa (n = 8) and lip (n = 2). The median patients' age at presentation was 34 (range: 16–40). Seven patients (5%) had Fanconi anemia with the gingiva being the most common location (4/7, 57%). All OSCCs were of keratinizing type. All cases tested for high-risk HPV (n = 34) were negative. On univariate analysis, high tumor budding was associated with decreased overall survival (OS) and distant metastasis free survival (DMFS), pattern of invasion correlated with OS and tumors with high stromal tumor infiltrating lymphocytes (sTIL) were associated with improved locoregional recurrence free survival (LRFS). Compared with patients 31 to 40-year-old, OSCC in the younger group was associated with significant less alcohol consumption (p = 0.011) and decreased DSS (p = 0.003) and DMFS (p = 0.023). On multivariate analysis, younger age (30 years or younger) was an independent prognostic factor for worse OS and DSS, whereas histologic grade was an independent prognostic factor for DSS. In summary, most OSCC in young patients occurred in non-smokers and did not occur in association with Fanconi anemia. Independent prognostic factors included age at presentation (30 years or younger) for OS and DSS, and histologic grade for DSS.

Introduction

Oral squamous cell carcinoma (OSCC), particularly of the oral tongue, was categorized as a distinct clinical entity in 1975 [1–3]. There has been a growing interest recently in OSCC in the young population (OSCCY) with multiple studies published on this subject [1–14]. This interest partially stems from a rise in the incidence of OSCCY [1, 15–21].

Although there has been no consensus on the age definition for OSCC in the young, most studies agreed on an age cutoff of 40 years [3]. OSCC is thought to have distinct pathophysiology in the younger population. While smoking and alcohol consumption constitute major risk factors of OSCC in the older population, they do not seem to play a major role in the pathogenesis of OSCCY [3, 4]. Chronic immunodeficiency states appear to be important factors in the carcinogenesis of OSCCY. For instance, Fanconi anemia (FA) patients have a high-risk of developing head and neck cancer, estimated to be 40% by the fourth to sixth decade of life [3, 22]. Furthermore, while human papillomavirus (HPV) infection and p53 mutation could promote the pathogenesis of oropharyngeal SCC, their role in OSCCY has not been well demonstrated [3, 4].

The reported outcome of OSCCY varied among studies. Multiple authors reported a worse survival rate in OSCCY as compared to the older population [19, 20, 23–25], while others showed a similar [13, 15] or even better survival in young patients [26]. Moreover, there is a limited reported data on the histopathologic prognostic parameters of OSCC in the young [26–30]. Only a few models have been developed to predict outcomes in OSCC, with most concurring on the significant of tumor invasive pattern and depth of invasion; however, the significance of these parameters has not be validated in the young population [28, 31–37]. Given the rarity of OSCC in the young population, most of the series described in the literature included a limited number of cases [3]. To further investigate the characteristics of OSCC in young patients, we performed a detailed clinicopathologic review of 150 cases in our institution and correlated the clinicopathologic parameters with the patients’ outcome.

Methods

Study Cohort and Clinicopathologic Review

The study was approved by the institutional review board of Memorial Sloan-Kettering Cancer Center. The pathology database was searched for OSCC diagnosed in patients who were 40-year-old or younger, who underwent surgical resection. A total of 150 patients fulfilling the selection criteria were included in this study (OSCC in 40-year-old or younger). Tumors of the oropharynx including the base of tongue were excluded from this study.

The following clinical and pathologic features were recorded: patients’ age, sex, smoking and alcohol consumption history, past medical history, tumor location, tumor greatest dimension, tumor thickness, depth of invasion (DOI) as defined by the American Joint Committee on Cancer (AJCC) 8th edition staging manual [38], lymphoid host response (LHR) and worst pattern of invasion (POI) as defined by Brandwein–Gensler et al. [36, 37], tumor budding as defined by the International Tumor Budding Consensus Conference (ITBCC) [39], stromal and intratumoral tumor infiltrating lymphocyte percentage (STil% and iTil%) as defined by the International Immuno-Oncology Biomarker Working Group [40], perineural invasion, lymphovascular invasion, margin status, AJCC pT stage, AJCC pN stage, administration of chemotherapy and/or radiation therapy.

Tumors were graded using a three-tiered system and classified as well, moderately or poorly differentiated based on the following criteria: Tumors with pushing invasion and minimal cytonuclear atypia were classified as well differentiated, tumors with infiltrative growth, large tumor islands, and moderate cytonuclear atypia as moderately differentiated, and tumors with small clusters of tumor cells and marked cytonuclear atypia were classified as poorly differentiated.

DOI was measured from an imaginary horizontal line stretching from the basement membrane of the adjacent normal squamous epithelium to the deepest tumor cluster. Worst POI was categorized into five patterns: POI 1 – broad pushing tumor front; POI 2—finger-like pushing invasion; POI 3—large tumor islands > 15 cells; POI 4—small tumor island ≤ 15 cells or tumor strands; and POI 5 – satellite tumor nodule at least 1 mm away from the main tumor [36, 37]. LHR was classified as: LHR 1: continuous rim of lymphoid tissue at the edge; LHR 2: discontinuous patch of lymphoid infiltrate; and LHR 3: no response [36, 37]. The number of tumor buds (defined as a tumor cluster of < 5 tumor cells) was counted using a 20 × objective at the hotspot. Tumor budding was then classified as low: 0–4 buds; intermediate: 5–9 buds; and high: ≥ 10 buds [39]. STil% and iTil% were defined as the percentage of stromal and intratumoral area covered by mononucleated inflammatory cells (lymphocytes and plasma cells) [40]. Any history of smoking was considered relevant. Alcohol intake was considered significant when reported beyond the level of occasional social consumption.

Statistics

All statistical analyses were performed using the SPSS software 24.0 (IBM Corporation, New York, NY, U.S.). The primary outcomes were overall survival (OS) and disease specific survival (DSS). Distant metastasis free survival (DMFS) and locoregional recurrence free survival (LRRFS) were also calculated. The prognostic significance of each parameter was calculated using log rank test. Factors significant on univariate analysis (except confounding factors due to selection bias being radiation therapy and chemotherapy) were subsequently subjected to multivariate analysis using Cox proportional hazards model. Hazard ratio (HR) and its 95% confidence interval (CI) were calculated. Additionally, Fisher’s exact test was performed to compare between patients with and without FA, as well as between patients ≤ 30-year-old and those between 31 and 40-year-old. For sTil%, ROC analysis was performed to determine the best cutoff value to predict the outcome. P values less than 0.05 were considered to be statistically significant.

Results

Clinicopathologic Characteristics of the Study Cohort

A total of 150 patients with OSCC diagnosed at the age of 40 years or younger were included in this study. For each patient, a median of 5 slides was reviewed (range 1–24). The median age at diagnosis was 34 years (range 16–40). The clinical and pathologic features are summarized in Table 1. There was a slight male predominance with a male to female ratio of 1.3:1. Among our cohort, 46 (31%) patients had OSCC at age of 30 or younger, whereas the remaining 104 patients were diagnosed between 31 and 40 years. The majority (n = 108, 74%) of the patients were Caucasian. Other ethnic groups included Indian (n = 12, 8%), Asian (other than Indian, n = 8, 6%) Hispanic (n = 8, 6%), Native American (n = 5, 3%), African American (n = 3, 2%), and Pacific Islander (n = 1, 1%). Merely 35% of the patients were smokers or ex-smokers, whereas the remainder of the cohort was composed of never-smokers. Twenty-eight (19%) patients reported significant alcohol consumption. There was an overlap between smoking tobacco and chewing tobacco/betel and smoking marijuana in some cases and all of these cases were included in the smokers category.

Table 1.

Clinicopathologic characteristics of the study cohort

	N	Column%
Age
≤ 30-year-old	46	31
31–40-year-old	104	69
Gender
Female	61	41
Male	89	59
Race
African American	3	2
American Indian	5	3
Asian	8	6
Hispanic	8	6
Indian	12	8
Pacific islander	1	1
Caucasian	108	74
Smoking
Non-smoker	97	65
Smoker/ex-smoker	52	35
Alcohol
Non-drinker/social	121	81
Drinker	28	19
Fanconi anemia
Absent	140	95
Present	7	5
Prior malignancy
Absent	135	92
Present	12	8
Autoimmune disease including hypothyroidism
Absent	126	86
Present	21	14
Allergy/Eczema/asthma
Absent	131	89
Present	16	11
Site of tumor
Buccal	8	5
Gingiva	9	6
Lip	2	1
Tongue	131	87
Tumor greatest dimension
≤ 2 cm	102	72
2.1–4 cm	38	27
> 4 cm	2	1
Associated dysplasia/in situ carcinoma
Absent	48	42
Mild	9	8
Moderate	53	47
Severe/SCC in situ	56	41
Verrucous hyperplasia	1	1
Present, NOS	3	3
Squamous cell carcinoma in situ
Absent	56	41
Present	81	59
Ulceration
Absent	34	33
Present	69	67
Histologic grade
Well-differentiated	5	3
Moderately-differentiated	107	74
Poorly-differentiated	32	22
Tumor thickness
≤ 5 mm	67	49
5.1–10 mm	34	25
> 10 mm	37	27
Depth of invasion (mm)
≤ 5 mm	60	50
5.1–10 mm	31	26
> 10 mm	29	24
Structure invaded
Lamina propria	24	17
Skeletal muscle	112	79
Bone	5	4
Worst pattern of invasion (POI)
POI 2	9	7
POI 3	65	52
POI 4	43	34
POI 5	9	7
Lymphovascular invasion
Absent	126	91
Present	13	9
Perineural invasion
Absent	86	62
Present	53	38
Lymphoid host response (LHR)
LHR 1	13	10
LHR 2	109	84
LHR 3	7	5
Stromal tumor infiltrating lymphocytes (sTil)%
< 15	34	27
≥ 15	90	73
Intratumoral tumor infiltrating lymphocytes (iTil)%
< 1	87	70
≥ 1	37	30
Tumor budding
Low	62	50
Intermediate	29	23
High	33	27
Margin status for in situ carcinoma
Negative	93	96
Positive	4	4
Margin status for invasive carcinoma
Negative	134	94
Positive	9	6
Radiotherapy
Absent	79	57
Present	59	43
Chemotherapy
Absent	104	77
Present	31	23
AJCC T stage
T1	78	53
T2	46	32
T3	13	9
T4	9	6
AJCC N stage
N0/Nx	100	68
N1	18	12
N2	20	14
N3	8	5
Extranodal extension
Absent	26	58
Present	19	42

In terms of past medical history, 7 patients (5%) had FA. Twelve patients had prior malignancy (lymphoma/leukemia in 6, sarcoma in 3, testicular germ cell tumor in 2, and basal cell carcinoma in 1). Sixteen patients (11%) had allergic conditions other than drug allergy, including seasonal allergy, allergic sinusitis, eczema and asthma, and 21 (14%) had a history of autoimmune disease, e.g. inflammatory bowel disease, lupus, hypothyroidism, (juvenile) rheumatoid arthritis, juvenile diabetes mellitus and Raynaud's phenomenon.

The most common site of the primary tumor was anterior/lateral tongue (n = 131, 87%), followed by gingiva (n = 9), buccal mucosa (n = 8), and lip (n = 2). Squamous cell carcinoma in situ was observed in 81/137 (59%) cases, whereas squamous dysplasia (mild or moderate) or verrucous hyperplasia was seen in 66/114 cases (60%). All tumors included in our study were classified as keratinizing SCC. Five tumors (3%) were graded as well differentiated, 107 (74%) as moderately differentiated and 32 (22%) as poorly differentiated (Fig. 1). Immunostain for p16 and/or in situ hybridization for high-risk human papillomavirus have been performed in 34 cases, all of which were negative. Lymphovascular invasion and perineural invasion were seen in 13 (9%) and 53 (38%) of cases respectively. Nine patients (6%) had a positive surgical margin for invasive squamous cell carcinoma, whereas 4 (4%) patients had a surgical margin involved by carcinoma in situ. The AJCC tumor (T) stage was T1 in 78 (53%), T2 in 46 (32%), T3 in 13 (9%) and T4 in 9 (6%). AJCC nodal (N) stage was N0/NX in 100 (68%), N1 in 18 (12%), N2 in 20 (14%) and N3 in 8 (5%). Among patients with lymph node metastasis, extranodal extension was seen in 19 (42%) cases.

Fig. 1.

Oral squamous cell carcinoma in young patients (YOSCC) (A-C): A: Well differentiated OSCC of the tongue; B: Moderately differentiated OSCC of the gingiva in a patient with Fanconi anemia; C: Poorly differentiated OSCC of the tongue

Patients with FA tended to have tumors originating from the gingiva (4/7, 57%) as compared to (5/135, 4%) patients without FA (Fisher’s exact test, p < 0.001). Other sites of primary tumors in patients with FA were oral tongue (n = 2) and lip (n = 1). The remaining clinical and pathologic features did not differ between patients with or without FA.

When comparing the group of 30 years or younger with those who were 31–40 years, the only factor that was significantly different was alcohol consumption status: 3/46 (7%) were drinker/ex-drinker in the lower age group, whereas 25/78 (32%) of 31 to 40-year-old patients consumed alcohol (Fisher’s exact test, p = 0.012). No other factors differed between the two groups.

Compared to patients with SCC of the oral tongue, patients with SCC of other oral cavity sites had a higher rate of second malignant primaries (4/19, 21% versus 6/131, 4.5%; Fisher’s exact test p = 0.024). Similarly, the rates of FA and T4 stage were higher in patients with SCC in sites other than the tongue (5/19, 26% and 7/19, 37% respectively) compared to patients with SCC in the oral tongue (1/131, 0.7% and 2/131, 1.5% respectively; Fisher’s exact test p < 0.001). The presence or absence of nodal involvement did not differ between these two groups.

Clinical Outcomes

A total of 147 patients had follow available up data. The median follow up period was 41 months (range 4–318 months). Eighteen patients developed distant metastasis. The site of metastasis in a descending frequency included bone (n = 13), lung (n = 9), brain (n = 2) and liver (n = 1). The p values of the univariate log rank test for various survivals (OS, DSS, DMFS and LRRFS) are shown in Table 2. The 5-year and 10-year OS were 83% and 78% respectively. The 5-year and 10-year DSS were 86% and 83% respectively. Patients who received radiation therapy or chemotherapy had decreased OS, DSS, DMFS and LRRFS, an association that can be explained by selection bias in which adjuvant therapies were given to patients with worse clinical manifestation. Younger age at the time of diagnosis was associated with decreased OS, DSS and DMFS (Fig. 2). Other factors that were associated with decreased DSS were: large tumor size, high histologic grade, Increase depth of invasion, presence of perineural invasion, positive margin, high AJCC pT stage and AJCC pN stage.

Table 2.

Univariate survival analysis using log rank test

	OS	DSS	DMFS	LRRFS
Age	0.006	0.003	0.023	NS
Gender	NS	NS	NS	NS
Smoking	NS	NS	NS	NS
Alcohol	NS	NS	NS	NS
Fanconi anemia	NS	0.028	NS	NS
Prior malignancy	NS	NS	NS	NS
Site of primary tumor	NS	NS	NS	NS
Histologic grade	< 0.001	0.003	NS	NS
Tumor greatest dimension	0.001	< 0.001	NS	NS
Tumor thickness	0.013	NS	NS	NS
Depth of invasion	0.003	0.011	NS	0.009
Structure invaded	NS	NS	NS	NS
Pattern of invasion	0.016	NS	NS	NS
Tumor budding	0.020	NS (0.068)	0.022	NS
Lymphovascular invasion	0.004	NS	NS	NS
Perineural invasion	< 0.001	0.001	0.009	NS
Lymphoid host response	NS	NS	NS	NS
sTIL%	NS	NS	NS	0.042
iTIL%	NS	NS	NS	NS
Margin for invasive carcinoma	0.010	0.042	NS	NS
AJCC pT stage	0.001	< 0.001	NS	NS
AJCC pN stage	0.026	0.018	NS	NS
Extranodal extension	NS	NS	NS	NS
Radiotherapy	< 0.001	< 0.001	0.012	0.017
Chemotherapy	< 0.001	< 0.001	0.007	< 0.001

OS overall survival. DSS: disease specific survival. DMFS distant metastasis free survival. LRRFS locoregional recurrence free survival. NS not significant

Fig. 2.

Kaplan Meier curves for survival

Tumor histologic grade predicted OS and DSS. High tumor budding was associated with decreased OS and DMFS. Pattern of invasion correlated with OS. ROC analysis was performed to determine the best cutoff of sTIL% to predict the outcome, which was determined at 15%. Tumors with high sTILs were associated with improved LRRFS but not OS, DSS and DMFS. sTIL status did not differ between smokers and non-smokers (p > 0.05). LHR and iTIL% did not predict clinical outcome.

The results of the multivariate analysis are shown in Table 3. Age (with a cutoff of 30 years or less) was an independent prognostic factor for OS (hazard ratio HR = 0.239, 95% confidence interval CI 0.064–0.895, p = 0.034) and DSS (HR = 0.136, 95% CI 0.033–0.560, p = 0.006), whereas histologic grade was an independent prognostic factor for DSS (HR = 6.753, 95% CI 1.345–33.912, p = 0.020). The other parameters did not reach significant levels.

Table 3.

Multivariate survival analysis using Cox proportional model

	Overall survival		Disease specific survival
	HR (95% CI)	P values	HR (95% CI)	P values
Age	0.239 (0.064–0.895)	0.034	0.136 (0.033–0.560)	0.006
Histologic grade	4.190 (0.971–18.075)	0.055	6.753 (1.345–33.912)	0.020
Perineural invasion	1.178 (0.218–6.358)	0.849	1.409 (0.214–9.269)	0.721
Margin status for invasive carcinoma	2.961 (0.493–17.794)	0.235	1.211 (0.182–8.036)	0.843
AJCC T stage	1.512 (0.579–3.947)	0.398	1.484 (0.573–3.841)	0.416
AJCC N stage	1.161 (0.576–2.340)	0.676	1.404 (0.753–2.615)	0.286
Depth of invasion	0.770 (0.093–6.394)	0.809	1.203 (0.395–3.665)	0.746
Tumor greatest dimension	1.023 (0.275–3.803)	0.972	1.124 (0.276–4.580)	0.870
Tumor thickness	1.508 (0.196–11.610)	0.693
Pattern of invasion	2.115 (0.797–5.617)	0.133
Lymphovascular invasion	1.808 (0.485–6.737)	0.378

Bold p values: significant p values. HR Hazard ratio, CI confidence interval

Discussion

Our cohort included 150 cases of OSCC in patients of 40 years of age or younger. Since there is no commonly accepted definition of “young age” in the context of oral SCC, the age limit for young patients has been arbitrarily assigned throughout the literature [3]. However, the majority of the previous reports of OSCCY considered a cutoff at 40 years of age, while others used different age categories (20 to 45) [3]. This difference in the age cutoff among literature reports might be one of the reasons leading to conflicting results. In addition to the age limit variability, many studies included only cases of oral tongue SCC or mixed OSCCY with other head and neck SCC [5, 41] making it challenging to draw definitive conclusions regarding OSCCY.

Most patients (63%) in our cohort were non-smokers. Furthermore, a small subset of patients (19%) reported alcohol consumption and there was significantly less alcohol consumption in patients younger than 30 years compared with patients 31 to 40-year-old (p = 0.011). Unlike OSCC in older patients in which smoking is the most important etiologic factor, smoking is not common among young patients with OSCC (reported range ≤ 5–51%) [2, 4–14], and there is no significant association between alcohol and OSCCY. This lack of association between OSCC and smoking or alcohol in the young population remains unclear, and could be attributed to a lower exposure time in the younger individuals. This suggests that young patients who develop OSCC without these conventional risk factors may represent a clinically distinct population with different disease pathophysiology and clinical behavior. In addition, similar to older population, while high risk HPV is strongly associated with oropharyngeal squamous cell carcinoma, it does not appear to play a significant pathogenic role in OSCC in young patients.

Prior studies have reported that FA patients have a higher risk for developing head and neck cancer [3] with oral cavity being the most common location, particularly oral tongue and gingiva [22]. In our study, only 5% of patients had FA. Furthermore, FA patients demonstrated particular characteristics: While the oral tongue was the most common location (87%) of all OSCCY in our cohort, only two (2/7, 29%) FA patients developed SCC of the oral tongue, and most (4/7, 57%) had SCC of the gingiva. It is also well known that carcinomas of the upper aerodigestive tract in young adults may be the first trigger for workup leading to the diagnosis of FA [22]. Therefore, our findings could reasonably imply that the suspicion of an underlying susceptibility disorder such as FA should be particularly high when a young patient presents with OSCC in a location different than the tongue.

In many of the literature series of OSCCY, the presence or absence of systemic or predisposition conditions (such as FA) has not been recorded [2, 4, 7, 9–14]. In our cohort, beside FA (5% of patients), 12 patients (8%) had prior malignancy, 17 (11%) had allergic conditions other than drug allergy, and 21 (14%) had a history of autoimmune disease including inflammatory bowel disease, lupus erythematosus, hypothyroidism, juvenile rheumatoid arthritis, juvenile diabetes mellitus and Raynaud's phenomenon. The relationship between immunocompromising microenvironments and developing cancer in young patients remains unclear. Since the presence of autoimmune/disimmune conditions might have an impact on the tumors’ immune microenvironment, we assessed the LHR and TILs in OSCCY. We found that tumors with high sTILs were associated with a decreased risk of local recurrence. While the presence of TILs has generally been associated with improved prognosis in Head and Neck SCC [42–44], the significance of TILs in OSCCY was not previously studied, and most reported studies have been limited by small cohort sizes, mixing of different anatomic sites and the inclusion of different age populations [40]. It would be very interesting to further investigate the tumors’ immune microenvironment and its prognostic significance, especially in regard to different subsets of T cells.

The clinical stage of OSCCY varied significantly among literature studies [4, 6–9, 11–14] and tumor size was not always recorded. In our study, tumor size and nodal stages were associated with worse DSS on univariate analysis, but they did not correlate with the outcome on the multivariate analysis.

Moreover, there has been a controversy regarding the reported prognosis of OSCCY of the oral tongue as compared to other locations in the oral cavity [20, 25, 45, 46]. Some authors found that patients with tongue SCC have a poorer prognosis than patients with SCC at other sites of the oral cavity [25] while others found a similar [46] or better outcome in oral tongue SCC as compared to SCC in other sites of the oral cavity [45]. In our cohort, the site of the primary tumor (including tongue vs non-tongue) did not correlate with the outcome.

Histologically, all OSCCs were of keratinizing type in this cohort. All tested cases for high-risk HPV/p16 were negative. Like in older population, high risk HPV seems to be rare and does not appear to play a significant pathogenic role in SCC of the oral cavity site in young patients. Similar to prior reported studies on OSCCY [2, 4–14], most tumors in this study showed a moderately differentiated histologic grade (74%) whereas 22% were poorly differentiated and only 3% were well differentiated. In terms of tumor pattern of invasion, we found POI 3 to be the most common pattern followed by POI 4. Previous studies have shown that histologic grading by differentiation is of limited prognostic value as compared to the POI [3] and that POI is a better predictor of local recurrence for head and neck SCC [37], particularly in the oral cavity [36]. However, most of these studies included OSCC of all ages [32, 36, 46, 47], and the significance of the POI has not been well established in the younger population. This reported absence of correlation between pathologic tumor grade and the outcome could be due, in part, to the subjectivity of histologic grading and the absence of a uniformly adopted grading system. In our study, we found that the grade of the tumor, as defined earlier, was an independent prognostic factor for OS and DSS, while POI was associated with OS on univariate analysis only. However, these results do not negate the importance of POI as a potential prognostic indicator. In fact, POI was incorporated within the histologic grading system that we used, practically by grouping POI 2 and PO 3 in the moderately differentiated category and including PO 1 and PO 4 in the well and poorly differentiated categories respectively.

Furthermore, while high tumor budding was recently identified as significant histologic event that correlates with poor prognosis and increased lymph node metastasis in SCC of the oral tongue in the general population, the prognostic significance of tumor budding has not been previously assessed in OSCCY [48–51]. In our series, high tumor budding was associated with an increased risk of distant metastasis and decreased OS and DMFS on univariate survival analysis, but not on multivariate analysis.

Several matched-pair analysis studies compared the outcome of oral SCC between the young and older populations, with controversial results [6, 15, 24, 52]. For instance, some studies showed worse prognosis in the younger population of less than 40 years in term of survival [52] or early local recurrence [24], while others found a similar [15] or a better prognosis at an age cutoff of 45 years [6, 53]. In our cohort, patients who were younger than 30 years had worse DSS compared to patients older than 30 years. Interestingly, only one patient with FA was among this younger age group, and he presented with a SCC of the lip. Previous studies have shown a very poor outcome in oral tongue SCC patients of 30 years of age or less: two studies reported the overall survival to be 45% and 55% respectively [7, 20], and another demonstrated a 100% death rate in recurrent cases [8]. Our cohort, which included the largest number of patients of < 30 years of age, showed however a much better outcome than the aforementioned studies (5-year and 10-year OS and 5-year and 10-year DSS of 83%, 78% and 86%, 83%, respectively). This difference in outcomes between our series and those of others could be explained by the fact that the prior reported studies included a smaller number of patients and were conducted in a different geographical location or in a remote time zone where recent therapeutic advances would not have been available.

It is important to mention some of the limitations of this study. While our cohort is, to our knowledge, the largest single institution series of OSCCY reported to date, a usual concern with cohorts from a large tertiary cancer center is the institutional and selection bias, where the population included may not always be representative of the general population. Furthermore, this study is retrospective, including patients managed over a long period of time, without a homogeneous standardized treatment approach throughout. Also, more or less aggressive therapeutic approaches might have been implemented differently in the younger patients [19] or in those with a susceptibility syndrome (such as FA) where the non-surgical treatments may be deleterious [22]. However, this is an inevitable limitation inherent to the nature of the population studied here.

In summary, our study showed that OSCCY most commonly occurred in non-smokers and does not always occur in the setting of known genetic predisposition syndromes. FA was found to be associated with OSCC in 5% of the cases with the gingiva being the most frequent site of the tumor. Young age (30 years or less) was found to be an independent prognostic factor for worse OS and DSS, while histologic grade was an independent prognostic factor for DSS. Additional studies of OSCCY are still needed for a better understanding of this disease in young patients, and to investigate its molecular and pathophysiologic mechanisms.

Funding

Research reported in this publication was supported in part by the Cancer Center Support Grant of the National Institutes of Health/National Cancer Institute under award number P30CA008748. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. The authors have disclosed that they have no significant relationships with, or financial interest in any commercial companies pertaining to this article.

Declarations

Conflict of interest

The authors do not have any conflicts of interest.

Ethical Approval

All procedures performed in this retrospective study involving human participants were approved by the Institutional Review Board (IRB) in our institution. The research did not involve animals. Informed consent can be provided if needed.