Clinicopathological and genetic analysis of primary intraosseous carcinoma not otherwise specified (PIOC NOS)

Abstract

Background

Primary intraosseous carcinoma not otherwise specified (PIOC NOS) is an extremely rare jawbone cancer, accounting for approximately 1–2% of all oral cancers. Considering its rarity, no established standard treatment exists for postoperative recurrence or metastasis. This study aimed to analyse the clinical manifestations of this disease and identify novel genomic abnormalities to aid identification of potential treatment strategies.

Methods

We examined the clinical information of 14 PIOC NOS cases treated at our institution in recent years and compared it with previous reports. Genomic analysis was conducted on seven formalin-fixed paraffin-embedded surgical specimens, assessing 523 DNA and 55 RNA cancer-related genes using next-generation sequencing.

Results

Our findings, along with previous literature, revealed the posterior mandible as the primary site in most cases, though this was rarely identified at the time of initial diagnosis. Recurrence within two years of surgery was common. Hotspot mutations in PIK3CA were observed in 60% of cases. Additionally, one patient had high TMB status.

Conclusions

Given the challenges in diagnosis and the potential for rapid recurrence or metastasis, early detection is crucial. Comprehensive genomic profiling is highly desirable, as genomic analysis may direct the development of targeted therapeutic agents to treat relapse.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12903-026-09040-9.

Introduction

The 2024 World Health Organization Classification of Head and Neck Tumours includes odontogenic carcinomas such as sclerosing odontogenic carcinoma, ameloblastic carcinoma, clear cell odontogenic carcinoma, ghost cell odontogenic carcinoma, primary intraosseous carcinoma not otherwise specified (PIOS NOS), odontogenic carcinosarcoma, and odontogenic sarcomas [1]. PIOC NOS denotes a central jaw odontogenic carcinoma that cannot be categorised as another type of odontogenic carcinoma. The classification was determined using the following criteria: PIOC NOS is a squamous cell carcinoma (SCC) originating within the jawbone. It is distinct from SCC of the oral mucosa as it does not have an early association with the oral mucosa during development. In X-ray images, gingival carcinoma frequently manifests as a flat-plate-shaped resorption, owing to its growth pattern from the outer to the inner region of the jaw bone. Conversely, PIOC NOS demonstrates significant resorption within the jawbone, frequently resulting in perforating resorption patterns that are indicative of jawbone destruction. Histopathologically, SCC was the primary subtype. Given its rarity, with a prevalence of less than one case per 100,000 people, the treatment approach for PIOC NOS is derived from the treatment modalities used for oral SCC.

Diagnosing PIOC NOS is challenging because of the absence of certain diagnostic criteria, such as oral mucosa continuity at onset, metastasis from another primary tumour, and association with other odontogenic or maxillary sinus tumours.

Next-generation sequencing (NGS) studies have revealed genetic alterations in head and neck SCC (HNSCC). Comprehensive research has identified significantly mutated genes, namely TP53, CDKN2A, FAT1, CASP8, AJUBA, PIK3CA, and NOTCH1, in HNSCC [2]. Recently, Yukimori et al. employed targeted NGS to analyse 50 genes in two cases of PIOC NOS [3]. TP53, KDR, and APC mutations were found in one case, as previously reported. Previously, the absence of NGS technology led to a lack of specific regulations concerning the demineralisation of surgical specimens, resulting in the common practice of acid demineralisation. Under conditions of strong acidic demineralisation combined with 10% formalin formic acid, the extraction of genomic DNA suitable for analysis proves challenging due to extensive fragmentation, with DNA fragments being reduced to approximately 50 base pairs as demineralisation time increases. In contrast, the use of 10% EDTA allows for the extraction of genomic DNA with significantly less fragmentation. It is also noteworthy that formalin-based warm fixation further contributes to DNA fragmentation. Furthermore, owing to the rarity of the disease, the utilisation of NGS-based analysis has only been documented in the above-mentioned report. Thus, little information is available regarding genetic abnormalities in PIOC NOS.

The primary aim of this study was to elucidate the clinical characteristics of PIOC NOS by systematically comparing clinical data from 14 recent cases treated at our institution with findings from previously published case reports and review articles. Additionally, we sought to identify genetic alterations associated with PIOC NOS through genomic analysis of formalin-fixed paraffin-embedded (FFPE) surgical specimens.

Materials and methods

This study was conducted with the approval of the Research Ethics Committee of The University of Osaka according to ethical guidelines for medical and health research involving human subjects (No. 937). Written informed consent was obtained for the use of all samples and clinical information in this study. All personal information was anonymized, and we ensured that patient privacy was fully protected.

Participants and samples

Of the 316 patients with oral cancer treated at the Second Department of Oral Surgery, The University of Osaka Dental Hospital between 2010 and 2023, 17 were clinically diagnosed with PIOC NOS, and all subsequently underwent surgical intervention. Of these, three patients were excluded following postoperative pathological assessment, which revealed a diagnosis of gingival carcinoma. The remaining 14 patients (10 males and 4 females) were confirmed to have a pathological diagnosis of PIOC NOS (Supplementary Figure S1). All patients underwent a comprehensive preoperative assessment, which included a thorough physical examination, panoramic X-ray, computed tomography (CT) scan, magnetic resonance imaging, positron emission tomography-computed tomography scan, ultrasound, and histopathology with biopsy technique. The diagnostic criteria for PIOC NOS included the initial absence of continuity with the oral mucosa, lack of metastasis from another primary tumour, and absence of association with other odontogenic tumours or tumours originating from the maxillary sinus.

Comprehensive clinical data, including smoking history, alcohol consumption, chief complaint, initial diagnosis, surgical procedure, lymph node metastasis, recurrence or metastasis, and prognosis, were collected. All the patients underwent surgical intervention (Fig. 1A, Table 1, and Supplementary Table S1).

Fig. 1.

Clinical characteristics of the 14 primary intraosseous carcinomas, not otherwise specified (PIOC NOS), in this cohort. A Demographic distribution by age and sex. The age range is broad, from 38 to 78 years, with a median of 67.5 years. However, most patients are older adults, with males accounting for over two-thirds of the population (71.4%). B The provided data displays the anatomical distribution of diagnosed PIOC NOS cases. The distribution of cases revealed that 3 (21.4%) were in the maxillary region and 11 (78.6%) in the mandibular region, with a notable predominance in the posterior mandibular area. C Disease-free survival and overall survival rates are presented in this cohort. Despite cases with short observation periods, the greatest number of recurrences and deaths occurred within 1 year of surgery. The x-axis represents time (months)

Table 1.

Clinicopathological characteristics of each patient with primary intraosseous carcinomas not otherwise specified (PIOC NOS) in this cohort

	Age	Sex	BI	Alcohol	Chief complaint	Site	Initial diagnosis	NAC	NAC effect	Tumor size (mm)	Treatment	Neck LN status	Recurrent/Metastasis	Prognosis (months)
1	49	M	450	Yes	Pain Mobility	Med-Maxi	SCC	PCE	Grd 1a	40	p-Maxill → add-p-Maxill	pN0	Yes	AWD (9.2)
2	74	F	0	No	Pain Swelling	Post-Lt-Mandi	SCC	PCE	Grd 1a	80	Lt-SMND + Lt-HemiMandi + Reconst	pN0	Yes	AWD (9.4)
3	64	M	396	Yes	Tongue discomfort	Post-Lt-Mandi	Ulcer	No	NA	90	Lt-ESOHND (SCM, SAN) + Lt-sMandi + Reconst	pN2a	No	NED (11.0)
4	70	F	0	No	Pain Swelling	Post-Lt-Mandi	Pericoronitis	No	NA	30	Lt-ESOHND + Lt-sMandi + Reconst	pN2b	Yes	AWD (12.2)
5	71	M	900	No	Swelling Dysesthesia	Post-Lt-Mandi	Alveolar osteitis	PCE	Grd 0	70	Lt-SMND + Lt-HemiMandi + Reconst	pN0	No	NED (13.6)
6	38	F	0	No	Pain Swelling Dysesthesia Inadequate healing after extraction	Post-Rt-Mandi	Pericoronitis	No	NA	30	Rt-SOHND + Rt-HemiMandi + Reconst	pN0	Yes	AWD (17.5)
7	71	M	0	No	Pain Dysesthesia Inadequate healing after extraction	Post-Rt-Mandi	Periodontitis	S1	NA	28	Rt-SOHND + Rt-HemiMandi + Rt-pMaxill + Reconst	pN2b	No	NED (21.6)
8	65	M	1600	Yes	Pain Swelling	Med- Mandi	GA	S1	NA	22	Bilat-SOHND + sMandi + Reconst	pN0	No	NED (23.5)
9	73	M	1200	No	Pain Swelling Mobility	Med-Maxi	Periodontal disease	No	NA	23	p-Maxill	NA	No	NED (23.9)
10	78	M	0	Yes	Pain Mobility Inadequate healing after extraction	Post-Rt-Mandi	Pericoronitis	No	NA	65	Rt-SOHND + Rt-HemiMandi + Reconst	pN0	No	NED (24.2)
11	46	M	400	Yes	Discomfort Inadequate healing after extraction	Post-Lt-Maxi	OMS	No	NA	22	Lt-ESOHND + Lt-pMaxill	pN0	No	NED (35.4)
12	62	F	0	NA	Pain Swelling Discomfort	Post-Rt-Mandi	Periodontal disease	S1 PCE	Grd 1a	38	Rt-ESOHND (SCM) + Lt-SMND + Rt-HemiMandi + Reconst	pN2b	Yes	DOD (3.2)
13	74	M	NA	Yes	Pain Swelling Dysesthesia	Post-Rt-Mandi	AA	S1	Grd 2B	60	Rt-ESOHND + Rt-sMandi + Reconst	pN0	No	NED (105.6)
14	60	M	NA	NA	Pain Swelling Dysesthesia	Post-Rt-Mandi	Mandibular cyst	S1	Grd 3	40	Rt-ESOHND + Rt-sMandi + Reconst	pN0	No	NED (114.9)

Abbreviations: BI Brinkman Index, NAC neoadjuvant chemotherapy, PCE paclitaxel + carboplatin + cetuximab, NAC Neoadjuvant chemotherapy, Post posterior (distal to canine), Med median (incisor to canine), Rt right, Lt left, Bilat bilateral, Maxi maxilla, Mandi mandible, SCC squamous cell carcinoma, GA gingival abscess, OMS odontogenic maxillary sinusitis, AA alveolar abscess, AWD alive with disease, NED no evidence of disease, DOD dead of disease, RT radiotherapy, CTx chemotherapy, Maxill maxillectomy, Mandi Mandibulectomy, p partial, s segmental, add Additional, Reconst Reconstruction, Ext extraction, SMND submandibular neck dissection, ESOHND extended supraomohyoid neck dissection, SOHND supraomohyoid neck dissection, SCM sternocleidomastoid muscle, SAN spinal accessory nerve, IJV internal jugular vein

Targeted NGS

NGS analysis was conducted on seven FFPE surgical specimen samples, focusing on genes associated with solid tumours (DNA, 523 genes; RNA, 55 genes) (DNA, 1.95 Mb; RNA, 358 kb). Nucleic acids were extracted from FFPE sections measuring 5 μm, targeting an area of at least 800 mm². NGS was used to analyse the coding regions of each target gene.

Analysis of PIOC NOS samples often necessitates demineralisation due to the presence of hard tissues such as bone and teeth. Among the 14 confirmed cases, 7 underwent acid-based demineralisation. However, acid demineralisation leads to significant nucleic acid fragmentation, rendering high-quality DNA and RNA extraction for panel sequencing virtually infeasible; thus, these cases were excluded from molecular analysis. Of the remaining seven cases, two were excluded due to insufficient recovery of nucleic acids, which precluded successful library preparation. The TruSight Oncology 500 workflow (Illumina, San Diego, CA, USA) was employed for target enrichment using capture methods. Library preparation was performed using the hybrid capture-based TruSight Oncology 500 Library Preparation Kit (Illumina), and libraries were sequenced using the NextSeq 550Dx (Illumina) at a 101 bp, paired-end.

The sequencing data encompassed clinically relevant classes of genomic alterations, including single-nucleotide variants (SNVs), small insertions and deletions (indels), gene amplifications, tumour mutational burden (TMB), and microsatellite instability (MSI) for DNA, as well as gene fusions and splice variants for RNA. SNVs and indels were identified in tumour DNA samples and were reported when they had an allele frequency > 5% and a sequencing depth > 100 ×. Variants with a frequency > 1% in publicly available SNP databases (1000 Genomes Project, Tohoku Medical Megabank Organization 8.3KJPN, Exome Aggregation Consortium) and synonymous and non-coding variants were excluded. Amplifications were reported when the copy number was ≥ 6, assuming 100% tumour content. However, for key oncogenes such as ERBB2, EGFR, FGFR1, FGFR2, FGFR3, FGF3, FGF4, MYC, MYCL, and MET, amplifications were reported from a threshold of ≥ 4 copies.

Fusion genes were identified by analyzing transcript sequences from tumour RNA samples. All fusions arising from chromosomal translocations, insertions, or inversions were included in the report. Splice variants were evaluated by examining transcript sequences from tumour RNA samples, and known splicing abnormalities in AR, EGFR, and MET were reported. TMB was calculated as the number of somatic variants per megabase within the coding regions of the target genes. For MSI assessment, instability was defined as ≥ 20% of microsatellite regions displaying variation. The clinical significance of each variant was interpreted using public databases, including ClinVar, BRCA Exchange, and the IARC TP53 Database. Variants associated with approved therapeutic agents were reported if they fulfilled the eligibility criteria for indication. Additionally, relevant clinical trial information was curated by querying major clinical trial databases (UMIN, JAPIC, JMACCT, jRCT, ClinicalTrials.gov). Gene variants were reported when they matched the inclusion criteria of an active study, based on the trial’s objectives, summary, and eligibility requirements.

Results

Clinicopathologic characteristics

The clinical features of the 14 patients are shown in Fig. 1, Table 1, and Supplementary Table S1. Of the total patients, 10 (71.4%) were male and 4 (28.6%) were female. The age at diagnosis ranged from 38 to 78 years, with a median age of 67.5 (Fig. 1A). The main symptoms were pain and swelling (Table 1). Regarding the distribution, 21.4% of the cases were located in the maxilla, with the majority (78.6%) occurring in the mandible (Fig. 1B). Among the 14 cases, two were observed in the midline and one in the left posterior region of the maxilla. Additionally, one case was found in the midline, four in the left posterior region, and six in the right posterior region of the mandible. Most cases (10/14, 71.4%) occurred in the posterior region of the mandible, as shown in Fig. 1B. Case 11 had a documented history of a radicular (root) cyst, and pathological evaluation in cases 8, 9, 13, and 14 revealed the presence of cystic walls. Case 12 was classified as solid type, while case 14 was considered to have originated from an odontogenic cyst (see Supplementary Table S1 for details). No patient had a family history of oral cancer. Recurrences, metastases, and mortality events occurred within one year following surgery (Fig. lementary Table S1C).

Only two patients (14.3%) were initially diagnosed with PIOC NOS, and obtaining a diagnosis proved to be a considerable challenge. Preoperative chemotherapy was administered to eight patients, with varying levels of efficacy observed (grade 0, one case; grade 1a, three cases; grade 2, one case; grade 3, one case; and unknown, two cases). In two patients, the surgical procedure was limited to primary tumour resection. Primary tumour resection and cervical lymph node dissection were performed in another patient, whereas 11 patients underwent primary tumour resection, cervical lymph node dissection, and reconstruction.

Following surgery, five patients (35.7%) experienced recurrence or metastasis, with two patients (14.3%) developing distant metastasis within a year. Additionally, six patients were diagnosed with a relatively small PIOC NOS with a specific tumour diameter of 30 mm. Despite the small Tumour size, two of the six patients experienced recurrence or metastasis.

Owing to the rarity of PIOC NOS, the literature is predominantly composed of case reports [4–29]. Additionally, several review articles [30–33] have succinctly summarised these studies. A PubMed search using “primary intraosseous squamous cell carcinoma” as the keyword yielded 31 articles (Supplementary Table S2) [3–33]. Following a meticulous examination of the References section and the removal of duplicate instances, 434 cases were documented in English papers. A clinicopathologic meta-analysis was conducted by integrating previous reports and our case studies (Table 2). Although there were numerous missing values, the available data indicate a higher prevalence among male patients. No clear association was observed with smoking or alcohol consumption. The primary site of tumour occurrence was predominantly the mandible, and the frequency of lymph node metastasis appeared to be low.

Table 2.

A meta-analysis of the clinicopathological features of each patient diagnosed with primary intraosseous carcinoma not otherwise specified (PIOC NOS)

Characteristic	All patients (n = 448)
	No. of patients	%
Sex
Male	303	67.6
Female	144	32.2
N/A	1	0.2
Age (years old)
5–60	73	16.3
> 60	42	9.4
N/A	333	74.3
Smoking
Yes	24	5.4
No	40	8.9
N/A	387	85.7
Alcohol
Yes	13	2.9
No	48	10.7
N/A	387	86.4
Tumor site
Maxilla	65	14.5
Mandible	383	85.5
Diagnosis at the first visit
Squamous cell carcinoma (SCC)	5	1.1
Non-SCC	34	7.6
N/A	409	81.3
Tumor type
Odontogenic cyst(s)	142	31.7
Non odontogenic cyst(s)	102	22.8
N/A	204	45.5
Lymph node metastasis
Yes	64	14.3
No	316	70.5
N/A	68	15.2

Smoking, alcohol consumption, and oral cancer are well-established links. Nevertheless, no association between these factors and PIOC NOS was detected in this cohort or previous studies, suggesting that the aetiology of PIOC NOS differs from that of common oral cancers, such as tongue cancer.

Targeted NGS

Mutations in PIK3CA were identified in three of the five cases (60%), followed by CDKN2A and TP53 in two cases (Fig. 2, Table 3). The PIK3CA mutations identified in these three cases were located in hotspots (p.E542K, p.E545Q, and p.K111E), recognised as driver mutants that enhance the kinase activity of PIK3CA. Among the three cases harboring PIK3CA mutations, one patient experienced recurrence at 5.9 months post-treatment, and another developed metastasis within 1 month. Of the two cases with TP53 mutations, one exhibited metastasis at 1 month, while the other presented with both recurrence and metastasis at 3.2 months. Notably, one case harbored concurrent mutations in both PIK3CA and TP53. All instances of recurrence and metastasis occurred within the first year following treatment. In one case, the TMB result was high (> 10 mutations/Mb). While the study is primarily exploratory, the identification of disease relapse and high TMB status in Case No. 2 through targeted NGS provides indisputable evidence supporting the potential effectiveness of an immune checkpoint inhibitor. In this case, the diagnosis of TMB-high was exclusively used. When a test reimbursed as a companion diagnostic agent confirms a TMB-high diagnosis, the administration of an immune checkpoint inhibitor can be considered.

Fig. 2.

The consequences of genetic changes using targeted gene sequencing. The genes exhibiting the highest number of genetic abnormalities were arranged in descending order. Mutations in PIK3CA were detected in three of the five cases, specifically in the hotspots

Table 3.

Results of targeted next-generation sequencing of cancer-related genes in seven PIOC NOS cases

Case	Tumor purity	MS-status	TMB	Genomic findings	Approved drugs (PMDA)	Clinical trial (Japan)	Approved drugs (FDA)
1	80%	2/124 (1.6%)	7.8/Mb	PIK3CA p.E545Q hotspot missense c.1633G > C, 103/443 (23.3%)	NA	NA	alpelisib + fulvestrant (PIK3CA) capivasertib + fulvestrant (PIK3CA)
2	70%	6/120 (5.0%)	16.4/Mb	CDKN2A p.E120* nonsense c.358G > T, 314/1100(28.5%) MDM2 amplification Copy number: 10.822 NOTCH1 p.C387* nonsense c.1160_1161delinsAA, 803/1915 (41.9%) NOTCH2 p.Q1791* nonsense c.5371C > T, 362/1902 (19.0%) PIK3CA p.K111E hotspot missense c.331A > G, 99/561 (17.6%) TP53 p.N200fs* frameshift indel c.599del, 794/1903 (41.7%)	Pembrolizumab (TMB)	NA	Pembrolizumab (TMB) alpelisib + fulvestrant (PIK3CA) capivasertib + fulvestrant (PIK3CA)
3	40%	Testing for genomics not possible
4	70%	2/121 (1.7%)	5.5/Mb	CCND1 amplification Copy number: 10.952 FGF19 amplification Copy number: 10.742 FGF3 amplification Copy number: 10.084 FGF4 amplification Copy number: 9.852 TP53 p.R342* nonsense c.1024C > T, 374/594 (63.1%)	NA	PF-07220060 (CCND1) NCT04557449 (CCND1) jRCT2031230121 (CCND1)	NA
5	70%	Testing for genomics not possible
11	30%	3/123 (2.4%)	7.0/Mb	CASP8 p.L385* nonsense c.1154 T > G, 109/1151 (9.5%) FAT1 p.D850fs* frameshift indel c.2547_2548ins, 147/2162 (6.8%) NCOR1 p.R363* nonsense c.108C > T, 34/452 (7.5%) PIK3CA p.E542K hotspot missense c.1624G > A, 27/488 (5.5%)	NA	NA	alpelisib + fulvestrant (PIK3CA) capivasertib + fulvestrant (PIK3CA)
14	20%	2/117 (1.7%)	4.7/Mb	CDKN2A p.D84N hotspot missense c.250G > A, 125/1166 (10.7%)	NA	NA	NA

Abbreviations: MS microsatellite, TMB tumor mutation burden, PMDA Pharmaceuticals and Medical Devices Agency, FDA Food and Drug Administration, NA not available

Gene names are shown in bold for clarity

Case pertaining to PIOC NOS with TMB-high

Figure 3 presents a case (Case No.2) of PIOC NOS marked by a high TMB (16.4/Mb). The patient displayed swelling on the left side of the mandible and paralysis of the left inferior alveolar nerve. Upon examination of the intraoral cavity, a lesion was identified that manifested as an ulceration on the gingival mucosal surface on the left mandible (Fig. 3A). Contrast-enhanced CT scans revealed a 70 mm lesion in the left mandibular molar, characterised by a heterogeneous centre (Fig. 3B). FDG-PET imaging further revealed a 50 mm mass with bone infiltration in the left mandible and mandibular angle, showing pronounced FDG accumulation (Fig. 3C). No lymph node involvement or distant metastases was detected. A biopsy confirmed a histopathological diagnosis of SCC. The patient was administered a preoperative chemotherapy regimen known as PCE therapy, which included paclitaxel, carboplatin, and cetuximab. A noticeable reduction in tumour size was observed after chemotherapy (Fig. 3D). Surgical intervention (left submandibular neck dissection [levels IA and IB], left hemimandibulectomy, plate fixation, and pectoralis major musculocutaneous flap) was performed, confirming the diagnosis of PIOC NOS in the left mandible through histopathological examination (Fig. 3E, F).

Fig. 3.

A patient with PIOC NOS (Case 2) representative displaying external and intraoral photographs, radiological images, and macroscopic and microscopic features. A Photographs taken externally and intraorally during the initial examination. Observations revealed swelling and palsy of the inferior alveolar nerve on the left mandible. The blue circle shows intraoral lesions. Edematous and ulcerative lesions were noted on the gingival mucosa, spanning from the left mandibular molar region to the posterior molar region. Following the tissue biopsy, the patient received a diagnosis of squamous cell carcinoma and subsequently underwent preoperative chemotherapy. B A transverse view of the body axis captured through a contrast-enhanced CT scan before the initiation of preoperative chemotherapy. The lesion has been encircled using the colour red. The lesion measuring 70 × 54 × 40 mm, displaying a heterogeneous contrast effect, was observed to be centred on the left mandibular molar. The lesion and the left submandibular gland were in proximity, with an indistinct boundary. The cortical bone in the buccolingual region displayed partial discontinuity. C FDG-PET imaging was performed before preoperative chemotherapy. The axial cross-sectional image of the body revealed a lesion measuring 50 mm in diameter, located in the left mandible and mandibular angle, exhibiting significant bone invasion. Additionally, there was a high FDG uptake (SUVmax: 11.38). The absence of metastasis to the lymph nodes or distant sites was noted. D Post-chemotherapy FDG-PET imaging. A residual lesion was observed on the left side of the mandible, demonstrating a decrease in size but remaining significantly concentrated (SUVmax, 16.61). No lymph node metastasis or distant metastasis was observed. E Gross image of the surgical specimen after formalin fixation. The centre of the section shows a white, hard, relatively homogeneous lesion. F Microscopic image using hematoxylin and eosin stain. The histopathological examination of the surgical specimen indicated an oral squamous cell carcinoma exhibiting keratinisation and a distinctive paving stone-like growth pattern characterised by the formation of small and large nests. The tumour displayed progressive growth by absorbing and eroding bone, with tumour cells detected close to the buccal and lingual transection. The cytoplasm exhibited acidophilic degeneration and vacuolar degeneration. It is plausible that the degenerative changes resulted from the effects of preoperative chemotherapy. (black bar: upper panel, 500 µm; lower panel, 100 µm). G Cross-sectional view of the body axis using postoperative contrast-enhanced CT. A single metastatic lymph node measuring 17 × 13 mm with internal necrosis was observed in the left upper internal deep cervical lymph node (circled in red)

Subsequent FDG-PET imaging post-surgery showed an enlarged lymph node in the left cervical region. Additionally, postsurgical contrast-enhanced CT imaging detected a metastatic lymph node in the left upper cervical region, featuring an internal necrotic area measuring 17 mm × 13 mm (Fig. 3G). The patient underwent left selective neck dissection (Level IIA-V) to address the recurrent lesion after surgery.

Discussion

This study is one of the most comprehensive reports (14 cases) from a single institution. Oral squamous cell carcinoma (OSCC), the most common form of oral cancer, is more prevalent in males and most frequently involves the tongue (25.4%), labial/buccal mucosa (21.7%), and gingiva (14.0%). In contrast, within our cohort, PIOC NOS also showed a male predominance, but was primarily localised to the posterior mandible. This distinct anatomical distribution may reflect differences in etiological factors, particularly the relevance of smoking and alcohol consumption as potential risk factors [34]. Notably, we were able to conduct targeted NGS analysis on seven cases of PIOC NOS, detecting genetic abnormalities in five cases. Out of the five patients, three exhibited hotspot mutations in the PIK3CA gene, while one patient displayed high TMB.

According to recent research by Yukimori et al., targeted NGS was utilized for the first time in two cases of PIOC NOS [3]; one patient did not exhibit any of the identified mutations, whereas the other presented with mutations in TP53, KDR, and APC. A comprehensive study identified TP53, CDKN2A, CASP8, PIK3CA, and NOTCH1 as significantly mutated genes in head and neck cancers, including oral cancer [35]. In our study, three of the five patients showed PIK3CA hotspot mutations. When compared to the 10–15% incidence rate of PIK3CA mutations in head and neck cancer [2], the substantial detection rate observed here indicates that PIK3CA could potentially function as a key driver gene in this disease. In the United States, two Food and Drug Administration-approved treatment protocols exist for patients with breast cancer PIK3CA mutations. The PI3K α-selective inhibitor alpelisib, combined with fulvestrant, showed greater efficacy in PIK3CA-mutant tumours in the SOLAR-1 (Clinical Studies of Alpelisib in Breast Cancer 1) trial [36] than endocrine therapy alone in advanced breast cancers [37]. Mutations found in PIOC NOS are specifically located in hotspot regions, indicating their potential efficacy. In a first-in-human phase Ia clinical trial (ClinicalTrials.gov NCT01219699), Jin et al. investigated alpelisib monotherapy in patients with solid tumors harboring PIK3CA mutations. Among 19 patients with HNSCC, the disease control rate (DCR) was 68.4%, and 7 out of 17 patients exhibited tumour shrinkage [38].

Additionally, one TMB-high case was detected. The discovery of a TMB-high case of this disease has implications for the use of immune checkpoint inhibitors. As only five cases underwent targeted NGS, the frequency of TMB-high in this disease remains unclear. The patient with a high TMB had a history of rheumatoid arthritis and was undergoing treatment with methotrexate. To our knowledge, the potential impact of methotrexate on somatic mutation burden has not been previously reported. Yu et al. demonstrated that cases of OSCC with high TMB may have increased responsiveness to immune checkpoint inhibitors [39]. Kobayashi et al. reported that in HNSCC, TP53 mutations were significantly more frequent in oral cancer and that TP53 status in recurrent tumours was associated with survival following recurrence or metastasis [40]. Within the current cohort, two patients with TP53 mutations exhibited metastasis, indicating the potential for high malignancy in PIOC NOS with TP53 mutations. As additional cases of PIOC NOS are studied and further genomic analyses are conducted, the utility of these mutations as predictive biomarkers may become evident, potentially informing therapeutic decision-making for this rare cancer subtype.

Although the majority of patients in this cohort (10 out of 14; 71.4%) were evaluated by their family dentist, only two had been accurately diagnosed at the initial visit, highlighting the diagnostic challenges associated with this condition. In our cohort, the disease was more prevalent in males and primarily affected the posterior mandible, compared with other oral cancers. The chief complaints of most patients were pain and swelling. No association between smoking, alcohol consumption, and PIOC NOS was detected in this cohort or previous studies, suggesting that the aetiology of PIOC NOS differs from that of common types of oral cancer, such as tongue cancer. Additionally, recurrence rates were high in cases in which recurrence occurred within one year of surgery. These characteristics align with prior Japanese and foreign literature on the PIOC NOS [41].

In Japan, the universal health insurance system provides broad access to advanced diagnostic imaging modalities, including CT, MRI, PET-CT, and ultrasonography. Consequently, comprehensive preoperative diagnostic evaluation is commonly performed in patients with suspected malignancies to assess disease extent and exclude alternative diagnoses. Although such an approach may differ from clinical practice in other countries, it reflects the characteristics of the Japanese healthcare system and routine oncologic management. While this approach may increase healthcare resource utilization, it may also contribute to accurate staging and treatment planning in diagnostically challenging diseases such as PIOC NOS.

The present study has some limitations. Although this study represents one of the largest series from a single institution, only 14 cases were included. Consequently, the findings may not fully capture the overall clinical and pathological characteristics of this rare disease. In addition, recognising the limitations of NGS analysis for PIOC NOS is necessary. Conducting genomic analysis in PIOC NOS is restricted by hard tissues such as bone and teeth, which often require demineralisation. Employing EDTA demineralisation is highly recommended as a proactive measure against acid demineralisation, especially for specimens intended for genomic analysis. However, even with demineralisation, a possibility of the NGS analysis being unsuccessful exists. Therefore, collecting freshly frozen samples is crucial for NGS, especially for sites lacking hard tissue. The collection of fresh-frozen tissue samples is essential for comprehensive genomic and transcriptomic analyses, such as whole-genome and whole-transcriptome sequencing, which are increasingly being promoted in Japan. However, the standardization of collection procedures and preservation management remains complex. To address this, the Japanese Society of Pathology has issued guidelines outlining best practices for the collection and storage of such specimens. As a result, the routine collection of fresh-frozen samples is gradually being implemented more systematically in clinical and research settings.

Imaging-based diagnosis of PIOC NOS presents significant challenges due to the wide variability in radiographic appearance, and timely follow-up within two years is crucial because of the high rate of recurrence within this timeframe. Lesions may range from well-defined, unilocular radiolucencies to ill-defined, diffusely infiltrative patterns. Notably, features commonly associated with slower-growing odontogenic lesions—such as tooth displacement and root resorption—have not been reported, potentially aiding the differential diagnosis. Histopathologically, PIOC NOS closely resembles conventional squamous cell carcinoma (SCC) of the oral mucosa; however, its underlying genetic alterations remain unknown. As the number of documented cases increases and genomic analysis progresses, the identification of disease-specific biomarkers may become possible. Such findings would facilitate earlier and more accurate diagnosis and support the development of precision medicine approaches for this rare malignancy. Comprehensive genomic profiling has the potential to identify therapeutic targets.

Biographies

Narikazu Uzawa

Is Professor of Oral and Maxillofacial Oncology and Surgery, Graduate School of Dentistry, The University of Osaka.

Shinichi Yachida

Is Professor of Cancer Genome Informatics, Graduate School of Medicine, The University of Osaka.

Authors’ contributions

NU and SYachida conceived and supervised the study. SYasuda, KO, and NU collected clinical and pathological data. SI, ST, YN, and YT performed genomic analyses. YN, YT, and SYachida analyzed and interpreted the data. SYasuda and SYachida drafted the manuscript. All authors critically reviewed the manuscript, read and approved the final version.

Funding

This work was supported by grants from Practical Research for Innovative Cancer Control from the Japan Agency for Medical Research and Development (AMED) (JP22ck0106546 to S. Yachida; JP25ck0106799 to S.Yachida); Project for Cancer Research and Therapeutic Evolution (P-CREATE) from AMED (JP21cm0106477 to S.Yachida); Project for Promotion of Cancer Research and Therapeutic Evolution (P-PROMOTE) from AMED (JP23ama221404 to S.Yachida; JP25ama221430 to S.Yachida); Core Research for Evolutional Science and Technology (AMED–CREST) from AMED (JP26gm2010009 to S.Yachida); AMED-ASPIRE form AMED (JP26jf0126022 to S.Yachida);United States-Japan Cooperative Medical Science Program from AMED (JP23jk0210009 to S.Yachida); JSPS KAKENHI (20H03662 to S.Yachida; 23H02892 to S.Yachida; JP25K02822 to N.U.); JST-AIP Acceleration Research (JPMKCR19U3 to S.Yachida); Integrated Frontier Research for Medical Science Division, Institute for Open and Transdisciplinary Research Initiatives, The University of Osaka (to S.Yachida); the Takeda Science Foundation (to S.Yachida); Project Mirai Cancer Research Grants (to S.Yachida); the Yasuda Medical Foundation (to S.Yachida); the Mitsubishi Foundation (to S.Yachida); the Princess Takamatsu Cancer Research Fund (to S.Yachida); and Yakult Bio-Science Foundation (to S.Yachida).

Data availability

All sequencing data generated in this study have been deposited in the DNA Data Bank of Japan (DDBJ) under accession number JGAS000837. Additional data supporting the findings of this study are available within the article and its supplementary information files.

Declarations

Ethics approval and consent to participate

This study was approved by the Research Ethics Committee of The University of Osaka in accordance with the ethical guidelines for medical and health research involving human subjects (Approval No. 937). Written informed consent was obtained from all participants for the use of their clinical information and tissue samples in this study.

Consent for publication

Written informed consent for publication was obtained from all participants included in this study.

Competing interests

The authors declare no competing interests.