Prognostic Value of the P16/P53 Immunohistochemical Association in Oropharyngeal Squamous Cell Carcinomas

Abstract

Importance:

Accurate HPV status diagnosis is essential for managing oropharyngeal squamous cell carcinoma.

Objective:

We aimed to refine HPV status determination using a cost-effective and reliable approach by integrating the p53 immunohistochemical marker with the conventional p16 marker. We assessed the association between HPV status and the p16/p53 tumor immunohistochemical status and analyzed patient survival based on p16/p53 expression.

Design, Setting, Participants, and Exposures:

This retrospective, single-center study included 285 patients with oropharyngeal squamous cell carcinoma (2017-2022) with known p16/p53 immunohistochemical profiles. In a subset of 92 patients, HPV status was formally determined using polymerase chain reaction or in situ hybridization.

Outcome Measures and Results:

The p16/p53 test was an excellent predictor of HPV status (specificity: 100%, sensitivity: 98.48%), significantly outperforming p16 alone (P < .01). Patients with a p16+/p53WT profile had superior 5-year overall survival (82.2%; 108/131) and recurrence-free survival (79.4%; 104/131). Patients with discordant profiles (p16+/p53MT, p16−/p53WT) had survival outcomes similar to HPV-independent oropharyngeal squamous cell carcinomas.

Conclusion:

The p16/p53 immunohistochemical profile appears to be a reliable surrogate marker for HPV status and a strong predictor of survival. It could help identify discordant cases, helping prevent inappropriate therapeutic de-escalation.

Relevance:

With the advancement of artificial intelligence, integrating additional markers into an algorithmic analysis could further refine patient classification and enable more personalized therapeutic approaches.

Graphical Abstract.

Key Message

• p16/p53 profile is a reliable marker for determining HPV status in patients with oropharyngeal squamous cell carcinomas. p16/p53 is a better predictor than p16 alone, with 98.73% diagnostic accuracy.

• Survival varies by p16/p53 profile, with p16+/p53WT patients having the best outcomes. p16+/p53MT defines an intermediate group, reclassifying some p16+ as non-HPV induced.

• The p16/p53 profile could help prevent inappropriate therapeutic de-escalation in certain patients.

Introduction

Since the early 2000s, the proportion of HPV-induced oropharyngeal squamous cell carcinomas (OPSCC) has been increasing in Western countries.^{1 -5} These HPV-induced OPSCCs represent a distinct pathological entity with better overall prognosis, and several clinical trials are currently ongoing to propose treatment de-escalation strategies for these HPV-induced OPSCCs.^{6 -8}

The gold standard for determining HPV involvement is the detection of HPV DNA or mRNA in tumor tissue, using polymerase chain reaction (PCR) or in situ hybridization (ISH). ⁶ These techniques are costly and not suitable for routine use, which is why the overexpression of the p16^INK4a protein (p16) through immunohistochemistry is used as a surrogate marker to determine HPV status in OPSCC. ⁹ Due to its reliability and ease of use, the p16 immunohistochemical profile is used as the sole marker for TNM classification (AJCC 8) of OPSCC, differentiating between p16-positive and p16-negative OPSCC.

However, the literature reports 5% to 20% false positives for the p16 marker, meaning p16-positive OPSCCs without detectable HPV DNA or mRNA by PCR.^{10 -14} In terms of prognosis, these p16-positive/HPV-negative OPSCCs have poorer survival outcomes compared to p16-positive/HPV-positive OPSCCs, with survival rates similar to HPV-unrelated OPSCCs.^15,16 These p16-positive/HPV-negative patients would be incorrectly classified as HPV-induced OPSCCs and could undergo inappropriate therapeutic de-escalation. ⁷

The tumor protein p53, encoded by the TP53 gene, is known to coordinate DNA repair processes. ¹⁷ Mutations in the TP53 are involved in HPV-unrelated OPSCCs, which are caused by the mutagenic effects of alcohol and tobacco.^18,19 Furthermore, there is a significant association between p53 immunohistochemical staining and the presence of TP53 mutations in OPSCCs, with a sensitivity of 96% and specificity of 100%. ¹⁹

The preliminary study by Benzerdjeb et al ²⁰ describes the p53 immunohistochemical staining had distinct immunoprofiles which were associated with wild-type TP53 (p53WT) or mutant-type TP53 (p53MT). It demonstrates that the p16-positive/p53WT immunohistochemical profile is a powerful predictor of HPV+ status in OPSCCs, and that the p16-positive/p53MT profile allows for the reclassification of certain p16-positive OPSCCs as HPV-unrelated. This method is easily applicable in routine practice. Although the prediction of HPV status based on the p16/p53 association has already been described in the literature, no studies have examined the prognosis of patients according to their p16/p53 profile.^{21 -25}

The objective of our study is to investigate the p16/p53 immunohistochemical association in patients with OPSCC and to analyze patient survival according to these profiles.

First, we analyzed a sample of patients with known HPV status (PCR or ISH) to confirm the association between the p16/p53 profile and HPV status. Second, we analyzed overall survival and recurrence-free survival in the entire OPSCC cohort based on their p16/p53 profile. Finally, in a complementary analysis, we described the effect of tobacco consumption on oncologic outcome according to p16/p53 status.

Materials and Methods

This was a retrospective, single-center cohort study conducted in our department including all consecutive patients diagnosed with OPSCC between January 2016 and December 2021. 285 patients were identified through the hospital’s electronic medical records, guaranteeing a sufficient number of patients with a minimum follow-up of 2 years. The sample size of 285 patients was determined based on the availability of complete p16/p53 immunohistochemical data and a minimum follow-up of 2 years. While no formal power analysis was conducted, this sample size is consistent with similar studies in the literature^20,26 and provides sufficient statistical power for survival analyses.

Inclusion criteria were: first occurrence of OPSCC, histology of squamous cell carcinoma, and oropharynx as the sole location.

Exclusion criteria were: history of upper aerodigestive tract squamous cell carcinoma, previous cervical irradiation, synchronous malignant lesions, or non-oropharyngeal locations. Following data were extracted from electronic medical records: patient’s characteristics such as date and age at inclusion, sex, tobacco consumption in pack-years, performance status (assessed using the Eastern Cooperative Oncology Group scale, ranging from 0 [fully active] to 4 [completely disabled]), date of death, date of recurrence, occurrence of a second primary tumor, initial treatment, treatment of recurrence, date of last follow-up. Tumor’s characteristics such as location, tumor size, TNM classification (7th edition, 2009), AJCC tumor stage (7th edition, 2009), p16 profile, p53 profile, and HPV status (ISH or PCR).

Patients were grouped based on their p16/p53 immunohistochemical profiles: p16+/p53WT, p16+/p53MT, p16−/p53WT, and p16−/p53MT. This classification was chosen to reflect the biological and prognostic heterogeneity of OPSCC (Figure 1).

Figure 1.

STROBE inclusion flow diagram.

For each patient, a single representative tumor block fixed in formalin and embedded in paraffin was selected for analysis. The p16 and p53 analyses have been systematically performed in our center since 2016.

P16/p53 Immunohistochemical Analysis

Based on the literature, ²⁰ the immunohistochemical analysis was performed using an automated immunostainer (Ventana BenchMark ULTRA; Roche, Tucson, AZ, USA). Immunohistochemical staining was performed on an automated immunostainer with a primary antibody, followed by the avidin-biotin-peroxidase complex method. Reactions were developed with diamino-3,30-benzidine tetrahydrochloride substrate solution (Sigmafast; Sigma-Aldrich, Tucson, AZ, USA). The tissues were counterstained with hematoxylin. A negative control was conducted using the same method but omitting the primary antibody. The primary antibodies and their final dilutions were as follows: anti-p53 (clone DO7; Ventana Medical Systems; pre-diluted by the manufacturer) and anti-p16 (clone E6H4; Ventana Medical Systems, Inc. Arizona, US.; pre-diluted by the manufacturer).

The overexpression of p16 staining (p16-positive) was considered as strong and diffuse nuclear and cytoplasmic staining of p16 in ≥70% of the tumor cells. The absence of p16 staining or <70% of the tumor was termed p16-negative.

Based on the literature,^26,27 two p53 immunohistochemical staining patterns were considered as “wild-type” patterns (p53WT): when heterogeneous nuclear staining of variable intensities was observed in tumor cells or in the basal and parabasal layers only.

The remaining 4 patterns were considered as “mutant-type” patterns (p53MT): (1) complete absence of staining in the tumor cells; (2) cytoplasmic staining with or without nuclear staining in the tumor cells; (3) uniform strong nuclear staining in at least 80% of the tumor cells; and (4) uniform strong nuclear staining in at least 80% of the basal and parabasal cells.

The validity of each procedure was verified by staining of internal or external controls. Internal controls included normal tonsillar tissue, and external controls were validated commercial slides.

The immunohistochemical study was analyzed independently by 2 pathologists (N.B., Z.A.). The interobserver variability between 2 pathologists for the p53 immunohistochemical staining patterns was Kappa = 0.87 [95% confidence interval (CI) = 0.77; 0.97], meaning almost perfect agreement. In the case of discordance between pathologists, slides were reviewed in order to reach agreement. All scorings were conducted blinded to clinical characteristics or outcome.

Data were extracted from a preliminary study of 92 patients ²⁰ in whom HPV tumor status was assessed using p16 immunohistochemistry, HPV DNA in situ hybridization, and HPV DNA polymerase chain reaction.

DNA ISH for high-risk HPV DNA was performed for p16-positive and p16-negative/p53WT cases. With this automated approach, 20% false negatives have been reported. ²⁸ For this reason, when DNA ISH was negative in these cases, DNA PCR was carried out to confirm the HPV status.

HPV DNA In-Situ Hybridization

5-µm sections from FFPE tumor blocks were analyzed for the presence of high-risk HPV DNA by ISH using an automated protocol and the Ventana HR HPV III probe set (Ventana Medical Systems). The detection of punctate signals localized to tumor nuclei was defined as a positive result (HPV+ tumor status). A case of OPSCC associated with HPV served as an external control for HPV+ tumor status. For cases that were p16-positive and p16-negative/WT-p53, DNA ISH for high-risk HPV DNA was performed (n = 63).

HPV DNA PCR

An area with >70% tumor cells was annotated for macrodissection. DNA was extracted using the QIAamp DNA FFPE tissue kit on a QIACube instrument (Qiagen -Venlo, Netherlands) and quantified with the Qubit fluorometer (Life Technologies, Carlsbad, California, US). For each test, eluates were diluted to a concentration of 5 to 10 ng/µl before conducting the PCR reaction in order to avoid its inhibition. In every reaction and assay, 5 µl (25-50 ng) eluate was used. For each batch of samples, a blank water sample was included in the extraction to monitor for cross-contamination. The blank sample was analyzed with the reference real-time PCR assay. HPV testing samples were tested by PCR using the CLART HPV2 PCR kit (Genomica, Madrid, Spain), according to the manufacturer’s recommendations. CLART HPV2 is approved for in-vitro diagnostic use in the European Union for the FFPE sample. The assay enabled the detection of 35 HPVs: 20 high-risk HPV and 15 low-risk HPV. This technology uses biotinylated primers (PGMY09/11) that amplify a fragment of 450 base pairs (bp) within the HPV L1 region. Co-amplification of a fragment of 892 bp in the region of the CFTR gene and a 1202 bp fragment of a transformed plasmid provides a control to ensure DNA extraction adequacy and PCR efficiency. Amplicons are detected by hybridization in a low-density microarray containing triplicate DNA probes specific to 35 types of HPV. PCR reactions were run on the PCR equipment Eppendorf Mastercycler ep Gradient S (Eppendorf AG, Hamburg, Germany) in 50-µl reactions containing 45 µl of reaction mix (Genomica, Madrid, Spain), 0.75 mmol/l MgCl (Thermo Fisher Scientific, Waltham, MA, USA) and approximately 50 ng of DNA. After a first denaturation step at 95°C for 5 minutes, reaction mixtures underwent 40 cycles at 94°C for 30 seconds followed by 55°C for 60 seconds, 72°C for 90 seconds and lastly a cycle of 4°C for 8 minutes. The genotyping results were analyzed and reported automatically on the Clinical Array Reader (Genomica, Madrid, Spain). All samples returning an invalid outcome were tested again, and the second outcome was considered definitive.

Statistical Analysis

Diagnostic performance measures (sensitivity, specificity, positive/negative predictive values, and accuracy) were calculated using standard formulas. Sensitivity = TP/(TP + FN), specificity = TN/(TN + FP), PPV = TP/(TP + FP), NPV = TN/(TN + FN), and accuracy = (TP + TN)/(TP + TN + FP + FN).

Kaplan–Meier curves were used to estimate overall survival (OS) and recurrence-free survival (RFS). OS was defined as the time from diagnosis to death from any cause, while RFS was defined as the time from diagnosis to loco-regional or metastatic recurrence or death. Patients alive without recurrence were censored at their last follow-up. The log-rank test was applied to compare survival distributions between groups. Association between p16/p53 profile and HPV status was assessed using Fisher’s exact test, as appropriate.

The Chi-square test and Wilcoxon rank-sum test were used for comparing categorical and continuous non-normally distributed variables, respectively. The biostatistician verified the correctness of these calculations and the interpretation of the results. He also confirmed that these tests were appropriate given the distribution of the data and the sample sizes.

A significance threshold of P < .05 was applied for all analyses.

Ethics

This study was approved by the national ethics committee (Number 2024-07-041-AL). Informed consent was obtained from all surviving patients via postal mail, with a clear option to refuse participation. For deceased patients, data were anonymized and used in accordance with national regulations on retrospective studies. All biological materials were handled according to the tumor biobank’s standardized procedures.

Results

Population Characteristics

We included 285 patients, categorized into 4 groups based on their p16/p53 profile: 131 patients with p16+/p53WT, 123 patients with p16−/p53MT, 12 patients with p16+/p53MT, and 19 patients with p16−/p53WT. The mean follow-up duration was 37.5 months. Their main characteristics are presented in Table 1, for descriptive purpose. Statistical testing for baseline imbalances was omitted due to the retrospective nature of the study and the lack of randomization.

Table 1.

Population Main Characteristics.

Variables	p16+/p53WT status (n = 131)	p16+/p53MT status (n = 12)	p16−/p53WT status (n = 19)	p16−/p53MT status (n = 123)	n
Age at diagnosis (mean [SD], years)	63.9 [9.98]	59.8 [8.61]	64.2 [9.32]	62.9 [9.23]	285
Sex
Female	30 (23%)	5 (38%)	6 (32%)	26 (21%)	67
Male	101 (77%)	7 (62%)	13 (68%)	97 (79%)	218
Tobacco consumption
No	46 (35%)	2 (15%)	1 (5.3%)	2 (1.6%)	51
Yes	85 (65%)	10 (83%)	18 (95%)	121 (98%)	234
Tobacco consumption (mean pack-years [SD])	15.7 [17.6]	25.8 [17.8]	37.6 [16.2]	39.4 [16.2]	285
Performance status
0	112 (85%)	9 (75%)	9 (47%)	53 (43%)	183
1	15 (12%)	3 (23%)	7 (37%)	51 (41%)	76
2	4 (3.1%)	0 (0%)	1 (5.3%)	17 (14%)	22
3	0 (0%)	0 (0%)	2 (11%)	1 (0.81%)	3
4	0 (0%)	0 (0%)	0 (0%)	1 (0.81%)	1
Tumor location
Lingual tonsil	33 (25%)	5 (38%)	5 (26%)	41 (33%)	84
Palatine tonsil	98 (75%)	7 (62%)	14 (74%)	82 (67%)	201
AJCC stage
1	9 (6.9%)	2 (17%)	0 (0%)	9 (7.4%)	20
2	12 (9.2%)	3 (25%)	4 (21%)	11 (9%)	30
3	53 (40%)	1 (8.3%)	1 (5.3%)	27 (22%)	82
4	57 (44%)	6 (50%)	14 (74%)	75 (61%)	152
Treatment
Surgery alone	7 (5.3%)	2 (16.6%)	2 (11%)	12 (9.71%)	23
Surgery + radiotherapy	50 (37.77%)	1 (8.3%)	4 (21%)	29 (23.81%)	84
Surgery + radiochemotherapy	36 (27%)	2 (17%)	3 (16%)	30 (24%)	71
Radiotherapy alone	8 (6.1%)	2 (17%)	0 (0%)	8 (6.5%)	18
Chemotherapy alone	4 (3.1%)	1 (8.3%)	3 (16%)	14 (11%)	22
Chemotherapy + radiotherapy	24 (18%)	4 (33%)	4 (21%)	19 (15%)	51
Chemotherapy + surgery	1 (0.76%)	0 (0%)	0 (0%)	2 (1.6%)	3
Chemotherapy + surgery + radiotherapy	0 (0%)	0 (0%)	2 (11%)	5 (4.1%)	7
Palliative care	0 (0%)	0 (0%)	0 (0%)	2 (1.6%)	2

Association Between the p16/p53 Status and HPV Status (Primary Outcome)

From this cohort of 285 patients, we identified 92 patients for whom the HPV status was formally determined using in situ hybridization +/− PCR. Among them, 26 patients (29%) were HPV-negative, while 66 patients (71%) were HPV-positive.

From the contingency table (Table 2), we compared the diagnostic performance of the p16/p53 test to the performance of the p16 test alone.

Table 2.

Contingency Table—HPV and p16/p53 Status.

Variables	HPV-negative status (n = 26)	HPV-positive status (n = 66)	n
p16/p53 status
p16+/p53WT	0 (0%)	65 (98%)	65
p16+/p53MT	9 (35%)	1 (1.5%)	10
p16−/p53WT	4 (15%)	0 (0%)	4
p16−/p53MT	13 (50%)	0 (0%)	13

Thus, the p16/p53 test demonstrates excellent specificity (100%) compared to the p16 test alone (60.71%), meaning that this test effectively eliminates false positives (ie, false “p16” cases). This is correlated with the test’s excellent positive predictive value (100%). The performance measures are described in Table 3.

Table 3.

Summary of p16/p53 Test Performance Measures (on a 92 Patients’ Sample).

Measure	p16/p53	P16 alone	P-value
Sensitivity	98.48%	100%	1.0
Specificity	100%	60.71%	.002
Positive predictive value	100%	85.3%	.004
Negative predictive value	96.30%	100%	1.0
Diagnostic accuracy	98.91%	88.04%	.012
Pearson correlation test	0.974	0.7198	<.01

Furthermore, the p16/p53 test exhibits excellent sensitivity (98.48%), correlated with an excellent negative predictive value (96.30%), with these values not being significantly different from those of the p16 test alone.

Additionally, the diagnostic accuracy (98.91%) of the p16/p53 test is significantly higher than that of the p16 test alone (88.04%). As a reminder, diagnostic accuracy indicates the total proportion of correct results (true positives and true negatives) relative to the total number of tested cases.

Finally, we found a significative association between HPV status and the p16/p53 profile using Fisher’s exact test, with a P-value of <.001. This indicates a very strong positive association between the results of the p16/p53 test and HPV status.

We can therefore conclude that the p16/p53 profile is an excellent predictor of HPV status. Based on this, we propose the following algorithm to determine HPV status (Figure 2).

Figure 2.

Proposed algorithm to determine HPV status based on p16/p53 profile.

Survival Analysis Based on the p16/p53 Profile (Secondary Outcome)

In this second part, the objective was to study patient survival according to their p16/p53 profile.

Analysis of Overall and RFS (at 2 and 5 Years)

Regarding the secondary outcome of our study, the overall survival rate for the entire cohort was 71.5% (204/285) at 2 years and 62.9% (179/285) at 5 years. A significant difference (P < .001) was observed between the different groups, with better survival at both 2 years (85.8%; 112/131) and 5 years (82.2%; 108/131) for patients with the p16+/p53 WT profile compared to the other groups. The Kaplan–Meier survival curves for 5-year overall survival are presented in Figure 3.

Figure 3.

Overall survival (a) and recurrence-free survival (b) according to p16/p53 status.

Regarding RFS (defined as loco-regional or metastatic recurrence, or death) for the entire cohort, it was 71.8% (205/285) at 2 years and 68.3% (195/285) at 5 years. A significant difference (P < .001) was also observed between patients with the p16+/p53 WT profile and the other groups, with RFS rates of 85.4% (112/131) at 2 years and 79.4% (104/131) at 5 years. The Kaplan–Meier curves for 5-year RFS are shown in Figure 3.

The percentages of overall and RFS at 2 and 5 years are presented in Table 4.

Table 4.

2- and 5-Year Overall and Recurrence-Free Survival According to p16/p53 Status.

Survival	Cohort (n = 285)	p16+/p53WT status (n = 131)	p16+/p53MT status (n = 12)	p16−/p53WT status (n = 19)	p16−/p53MT status (n = 123)	P
Overall survival
2-year (95% CI)	71.5% (66.4%; 77.0%)	85.8% (80.0%; 92.1%)	81.8% (61.9%; 100.0%)	63.2% (44.8%; 89.0%)	56.3% (48.1%; 66.0%)	<.01
5-year (95% CI)	62.9% (57.0%; 69.4%)	82.2% (75.3%; 89.6%)	68.2% (43.3%; 100.0%)	57.9% (39.5%; 85.0%)	44.1% (35.5%; 54.7%)	<.001
Recurrence-free survival
2-year (95% CI)	71.8% (66.4%; 77.5%)	85.4% (79.3%; 91.9%)	64.8% (42.1%; 99.8%)	54.0% (34.9%; 83.6%)	59.1% (50.2%; 69.5%)	<.01
5-year (95% CI)	68.3% (62.6%; 74.5%)	79.4% (72.0%; 87.6%)	64.8% (42.1%; 99.8%)	54.0% (34.9%; 83.6%)	57.8% (48.9%; 68.4%)	<.001

Regarding recurrence rates, they were 27.7% (79/285) for the entire cohort, 17.5% (23/131) for the p16+/p53 WT group, 33.3% (4/12) for the p16+/p53 MT group, 42.1% (8/19) for the p16−/p53 WT group, and 35.7% (44/123) for the p16−/p53 MT group.

Complementary Analyses

We conducted an analysis of overall and RFS stratified by tobacco consumption (yes/no) (Table 5). For this stratified analysis, in order to increase sample size and improve the statistical power of our tests, we compared the p16+/p53 WT group (considered HPV-positive) to a group combining the 3 other profiles—p16+/p53 MT, p16−/p53 WT, and p16−/p53 MT (considered HPV-negative).

Table 5.

2- and 5- Year Overall and Recurrence-Free Survival According to p16/p53 Status, Stratified by Tobacco Consumption.

Tobacco consumption	p16+/p53WT status (n = 131)		P	p16+/p53MT et p16− status (n = 154)		P
	Yes (n = 85)	No (n = 46)		Yes (n = 149)	No (n = 5)
Overall survival
2-year (95% CI)	85.2% (77.8%; 93.3%)	87.0% (77.7%; 97.3%)	.84	59.1% (51.6%; 67.7%)	60.0% (29.3%; 100.0%)	.89
5-year (95% CI)	81.9% (73.6%; 91.2%)	84.8% (75.0%; 95.8%)	.78	47.8% (39.8%; 57.4%)	—	—
Recurrence-free survival
2-year (95% CI)	82.2% (74.1%; 91.2%)	91.1% (83.1%; 99.8%)	.22	59.6% (51.6%; 68.9%)	40.0% (13.7%; 100.0%)	.25
5-year (95% CI)	76.9% (67.7%; 87.4%)	83.5% (71.6%; 97.4%)	.29	58.6% (50.5%; 68.0%)	40.0% (13.7%; 100.0%)	.25

The analysis of overall and RFS (at 2 and 5 years) stratified by tobacco consumption showed that tobacco use had no impact on survival, regardless of the group.

Discussion

The aim of our study was to confirm the concordance between HPV status determined by PCR/ISH and tumor immunohistochemical expression of p16/p53 in patients with OPSCC. Additionally, we sought to analyze patient survival based on their p16/p53 expression profile.

First, we demonstrated that the p16/p53 test has excellent diagnostic accuracy (98.91% vs 88.04% for the p16 test alone) and excellent specificity (100%). Our results indicate that the p16/p53 combination is a more robust predictor of HPV status compared to p16 expression alone, which, despite its high sensitivity, has a reduced specificity of 60.71%.

These findings are consistent with those of Benzerdjeb et al. ²⁰ Their retrospective study, conducted on 110 cases of head and neck squamous cell carcinoma (HNSCC), reported a diagnostic accuracy of 99% and a specificity of 100% for the p16/p53 test. The role of p16/p53 co-expression in HNSCC has also been investigated by Pakkanen et al, ²⁶ in a prospective study including 31 cases of HNSCC (not limited to oropharyngeal carcinomas). Their findings align with ours, showing a sensitivity of 83% and a specificity of 100%.

Next, we focused on patient survival according to their p16/p53 profiles. Patients with a p16+/p53WT profile exhibited better OS (85.8% at 2 years, 82.2% at 5 years) and RFS (85.4% at 2 years, 79.4% at 5 years) compared to patients with a p16+/p53MT profile (OS: 81.8% at 2 years, 68.2% at 5 years; RFS: 64.8% at 2 years, 64.8% at 5 years), a p16−/p53WT profile (OS: 63.2% at 2 years, 57.9% at 5 years; RFS: 54.0% at 2 years, 54.0% at 5 years), or a p16−/p53MT profile (OS: 56.3% at 2 years, 44.1% at 5 years; RFS: 59.1% at 2 years, 57.8% at 5 years).

We found a significant survival difference between patients with a p16+/p53WT profile (considered HPV-positive) and the other groups. Survival rates of discordant profiles (p16+/p53MT and p16−/p53WT) were similar to those of the p16−/p53MT group (considered HPV-negative), supporting the idea that this simple immunohistochemical technique is a promising surrogate marker for HPV status in OPSCC and could allow for the classification of discordant profiles as HPV-independent. However, given the retrospective design and lack of adjustment for confounders, these results should be validated in larger, prospective studies before clinical implementation.

Mehanna et al, ¹⁵ in their meta-analysis including 6882 patients with known p16 immunohistochemical profiles and HPV status determined by PCR or ISH, highlighted a similar issue with discordant p16/HPV profiles, which were associated with intermediate survival outcomes (5.4% of patients were p16+/HPV−, and 3.8% were p16−/HPV+). Their study demonstrated significant differences in 5-year overall survival, with excellent survival for p16+/HPV+ patients (81.1%) and poor survival outcomes for other groups, including discordant profiles (40.4% for p16−/HPV−, 53.2% for p16−/HPV+, and 67.9% for p16+/HPV−). The authors relied on expensive techniques to confirm HPV status.

Similarly, in the study by Pakkanen et al, ²⁶ the authors proposed an algorithm in which they recommended performing PCR or ISH-based HPV testing for discordant profiles (p16+/p53MT and p16−/p53WT). Our study provides a simpler and more reproducible predictive approach to identifying these intermediate profiles, which could facilitate personalized treatment strategies.

Additionally, we observed a clinical subset of smoking patients with tumors expressing p16. These patients are often treated as HPV-negative due to their smoking history and may not always be eligible for therapeutic de-escalation trials. We conducted an additional survival analysis stratified by smoking status (smoker vs non-smoker) and by cumulative tobacco exposure in pack-years (<10 pack-years vs ≥10 pack-years). These analyses demonstrated that overall and RFS at 2 and 5 years were not influenced by smoking status in the p16+/p53WT group. This finding suggests that the absence of p53 mutations may be associated with a purely HPV-induced carcinogenesis rather than the classical alcohol-tobacco-related carcinogenesis. Thus, smoking patients with a p16+ tumor and p53WT should be considered as HPV-induced and should receive appropriate HPV-driven treatment strategies.

To our knowledge, our study represents the largest series investigating the role of the p16/p53 combination as a surrogate marker for HPV status in OPSCC. Only one other study (Wang et al) ²² has examined patient survival based on p16/p53 profiles, but with a smaller cohort and without specifically focusing on OPSCC. Wang et al. also incorporated a third immunohistochemical marker (TLR9). Their study population included 85 cases of HNSCC. They reported that high p16 expression combined with mutated p53 expression serves as a reliable prognostic tool for survival in HNSCC patients. Their findings align with our results, showing that patients with a p16+/p53WT profile exhibit better overall and RFS compared to other studied profiles. Moreover, Wang et al suggested that adding TLR9 expression could further refine this classification.

The main strength of our study is the relatively large cohort size, which enhances the statistical robustness of our findings. Furthermore, our study proposes simple, reproducible, and reliable diagnostic criteria that could be easily integrated into clinical practice.

However, the primary limitation of our study is its retrospective nature, which introduces a potential confounding bias. Additionally, as a single-center study, the generalizability of our findings may be limited. Another limitation of our study is the lack of adjustment for potential confounders (eg, tobacco use, T/N stage, age) in survival analyses. While our findings suggest strong associations between p16/p53 profiles and survival, future prospective studies should include multivariate analyses to account for confounding factors. Immunohistochemical slide interpretation could also introduce classification bias, although this was controlled by a double reading in pathology. Additionally, the use of different HPV detection methods (ISH and PCR) may introduce variability in sensitivity and specificity. However, PCR was systematically performed when ISH results were negative in p16-positive or p16-negative/p53WT cases, minimizing the risk of misclassification. Future studies should ideally use a single, standardized method for HPV detection across all samples. Finally, systematic HPV status confirmation via PCR/ISH for all patients would have further strengthened the diagnostic performance assessment.

A larger prospective study incorporating systematic HPV status confirmation via PCR or ISH, along with the analysis of additional immunohistochemical markers, is needed to validate our findings. With the advancement of artificial intelligence, integrating additional markers (such as TLR9, as proposed by Wang et al ²² ) into an algorithmic analysis could further refine patient classification and enable more personalized therapeutic approaches.

Conclusion

In conclusion, the p16/p53 immunohistochemical profile appears to be a reliable surrogate marker for HPV status and a strong predictor of survival. While it helps identify discordant cases and may prevent inappropriate therapeutic de-escalation, these findings should be interpreted with caution due to the retrospective study design and limited subgroup sizes.

Further prospective validation is required to confirm these results. A larger-scale prospective study, in which HPV status is confirmed for all patients, would be valuable in validating the p16/p53 association. Additionally, studies on other immunohistochemical markers to further enhance the diagnostic power of this method would be beneficial.

With the advancement of artificial intelligence (AI), integrating additional immunohistochemical markers and algorithmic interpretation of a panel of biomarkers using AI could allow for better stratification of patient profiles, leading to more personalized treatment approaches.