Long-Term Survival Outcomes of HPV DNA–Negative Cervical Cancer Compared with HPV DNA–Positive Disease in a Thai Tertiary Cohort

Introduction

Cervical cancer remains a major public health concern globally. In Thailand, it was the sixth most common cancer in 2020, accounting for 4.8% of all cancer cases across the population.¹ The age-standardized incidence and mortality rates were 9.4 and 3.8 per 100,000 women, respectively.¹ Despite these figures, cervical cancer is considered a preventable disease, with effective screening methods such as HPV testing, Pap smear, or co-testing significantly reducing the disease burden.² Persistent infection with high-risk HPV types is a well-established etiological factor for the development of cervical cancer.³

The 2020 World Health Organization (WHO) Classification of Female Genital Tumors (5th edition) introduced a revised system categorizing cervical cancers into HPV-associated and HPV-independent types.^4,5 HPV-independent cervical cancer is associated with a worse prognosis and accounts for approximately 7% of squamous cell carcinomas and 15% of adenocarcinomas.⁵ As these subtypes cannot be reliably differentiated by morphological features alone, the WHO recommends ancillary testing using p16 immunohistochemistry and HPV DNA testing. Overexpression of p16 is a surrogate marker of HPV infection, as the viral E6 and E7 oncoproteins disrupt p53 and Rb tumor suppressor pathways, leading to upregulation of p16. In HPV-associated squamous cell carcinoma, diffuse p16 positivity supports the diagnosis.⁶

It is important to distinguish between HPV DNA–negative cervical cancer, as identified by molecular testing, and WHO-defined HPV-independent cervical cancer, which requires both HPV DNA testing and p16 immunohistochemical confirmation. Cases that are HPV DNA–negative by molecular assay alone may include false negatives due to low viral load, non-targeted genotypes, or assay limitations, and do not necessarily correspond to true WHO-defined HPV-independent tumors. Throughout this report, we use the terms “HPV DNA–negative” and “HPV DNA–positive” specifically to denote HPV DNA detection status, without implying WHO-based HPV-independent or HPV-associated classification.

Several recent studies have investigated the incidence and prognostic implications of HPV-independent cervical cancer. For example, Stolnicu et al reported an incidence of 11.2% for HPV-independent tumors in a multi-institutional, international study, with this subgroup having significantly lower recurrence-free survival (42.0% vs. 72.0%) than those with HPV-associated cancers.⁷ Meanwhile, Kaliff et al reported a 14.0% prevalence of HPV-independent tumors in a Swedish cohort, along with a markedly lower 5-year overall survival rate of 27.0%, compared to 69.0% in HPV-associated tumors.⁸ Similarly, Nicolás et al reported a 9.8% prevalence of HPV-independent tumors in a cohort from Barcelona, Spain, with significantly shorter disease-free survival (59.8 vs. 132.2 months) and overall survival (77.0 vs. 153.8 months) compared to HPV-associated cases.⁹ Using PANArray and Anyplex II assays, Chong et al reported an 18.5% prevalence of HPV-independent tumors in a Korean cohort, with a hazard ratio for disease recurrence of 3.97 (95% CI, 1.84–8.58; p = 0.0005) and a higher recurrence rate in HPV-negative patients (30.8% vs. 15.3%).¹⁰

Despite increasing recognition of HPV-independent cervical cancer in international cohorts, data from Southeast Asia remain limited. Most published series originate from Western European, North American, or East Asian populations, where HPV genotype distribution, screening practices, and healthcare access may differ substantially from Thailand and neighboring countries. The prevalence, clinicopathological characteristics, and survival outcomes of HPV DNA–negative cervical cancer in Thai patients have not been well characterized. Therefore, this study aimed to evaluate 5-year progression-free survival in HPV DNA–negative cervical cancer patients treated at a tertiary referral center in Thailand and to compare outcomes with HPV DNA–positive disease.

Materials and Methods

Study Design and Participants

The study protocol was approved by the Research Ethics Committee of Chulabhorn Royal Academy (IRB approval number: EC 011/2567) and was conducted in accordance with the ethical principles of the Declaration of Helsinki. Because this study involved retrospective review of existing medical records with minimal risk to participants and no direct patient contact, the requirement for written informed consent was waived by the ethics committee. All patient data were anonymized and handled confidentially prior to analysis. We included all patients diagnosed with cervical cancer of any stage between 2012 and 2014. Eligible patients had completed standard treatment and follow-up at Chulabhorn Hospital, with their complete medical records available. All cases were reclassified according to the 2018 International Federation of Gynecology and Obstetrics (FIGO) staging criteria for cervical cancer. Inclusion criteria were as follows: (a) histologically confirmed diagnosis of invasive cervical cancer with histologic subtype as squamous cell carcinoma, adenocarcinoma, or adenosquamous carcinoma; (b) HPV DNA testing performed using the HPV DNA genotyping test (Linear Array, Roche Molecular Systems); (c) clinical evaluation by a gynecologic oncologist; (d) availability of pretreatment imaging; and (e) receipt of complete standard primary treatment at Chulabhorn Hospital as described below. Exclusion criteria included (a) no evaluable response after completing treatment, (b) a history of previous malignancy (except non-melanoma skin cancer), or (c) prior hormonal therapy or chemotherapy before being diagnosed with cervical cancer. This study used anonymized routinely collected clinical data obtained during standard patient care.

Definition of HPV DNA–Positive and HPV DNA–Negative Cervical Cancer

Because p16 immunohistochemistry was not routinely performed at our institution during 2012–2014, HPV status in this study was defined based on HPV DNA detection alone. HPV DNA–positive cervical cancer was defined as detection of one or more HPV genotypes using the HPV DNA genotyping test (Linear Array, Roche Molecular Systems), whereas cases with no detectable HPV genotype were defined as HPV DNA–negative. Accordingly, HPV status in this report refers strictly to HPV DNA positivity/negativity and does not constitute definitive WHO HPV-associated/HPV-independent classification. The possibility of misclassification arising from false-negative HPV DNA results—due to low viral load, non-targeted genotypes, or sample quality—cannot be excluded. Given the absence of p16 immunohistochemistry, we refer throughout to groups as HPV DNA–positive and HPV DNA–negative.

HPV Genotyping and Standard Treatment Protocol

The HPV DNA genotyping test used in this study (Linear Array, Roche Molecular Systems), based on PCR amplification, detects 37 anogenital HPV types,^11,12 including all high-risk strains classified by the International Agency for Research on Cancer (IARC) 2012 report.¹³ High-risk HPV types included in the assay are classified as follows: Group 1 (carcinogenic)—HPV 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59; Group 2A (probably carcinogenic)—HPV 68; and Group 2B (possibly carcinogenic)—HPV 26, 53, 66, 67, 70, 73, 82. Treatment was administered according to the disease stage and institutional protocols. For early-stage disease, patients underwent radical hysterectomy with bilateral pelvic lymph node dissection. To treat locally advanced stages, definitive concurrent chemoradiotherapy (CCRT) was provided, consisting of external beam pelvic radiotherapy (EBRT) to a total dose of 50 Gy in 25 fractions over 5 weeks, combined with weekly cisplatin (40 mg/m²) or carboplatin (AUC 2) for patients with renal impairment. High-dose-rate intracavitary brachytherapy was also delivered at 7 Gy per fraction for four fractions for this patient group. For patients with bulky or metastatic disease, systemic chemotherapy was administered, typically cisplatin (50 mg/m²) or carboplatin (AUC 5) with paclitaxel (175 mg/m²) every 21 days. This chemotherapy was postponed until recovery in cases of hematologic, renal, or systemic toxicities.

Follow-up assessments included history-taking, pelvic examination, and symptom-directed evaluation. Radiological response was assessed 3 to 4 months post-treatment. Clinical follow-up visits were scheduled every 3 months for the first 3 years, then every 6 months until 5 years, and annually thereafter. Biopsy confirmation was obtained in all cases where recurrence was suspected.

Outcomes

The primary outcome was 5-year progression-free survival (PFS), defined as the time from completion of primary treatment to the first occurrence of locoregional recurrence, distant metastasis, or death from any cause. Given the median follow-up of 9.8 years, 10-year survival outcomes were additionally reported to provide a more complete description of long-term oncologic behavior in this cohort; these are presented as secondary exploratory analyses.

Secondary outcomes included the prevalence of HPV DNA–negative cervical cancer and 5-year overall survival (OS), defined as the time from treatment completion to death from any cause. Survival outcomes were evaluated for both HPV DNA–negative and HPV DNA–positive groups for comparative analysis.

Statistical Analysis

Sample size was estimated using a two-sample comparison of survival functions under exponential assumptions, with a two-sided α of 0.05 and 90% power, informed by hazard ratios reported previously. Based on the expected prevalence of HPV DNA–negative disease, the target cohort size was increased to allow sufficient representation and potential data loss. However, the final number of HPV DNA–negative cases (n=11) was lower than initially assumed and may have been insufficient to achieve the planned statistical power for between-group comparisons; consequently, all survival comparisons should be regarded as exploratory. The proportional hazards assumption for Cox regression was assessed using Schoenfeld residuals; no significant violations were identified. The number of events per variable in the Cox model was assessed and found acceptable for the unadjusted analyses presented. Continuous variables are presented as mean ± SD or median (interquartile range) and compared using the independent t-test or Mann–Whitney U-test, as appropriate. Categorical variables are presented as counts and percentages and compared using Fisher’s exact test. Kaplan–Meier methods were used to estimate PFS and OS, with between-group comparisons performed using the Log rank test. Cox proportional hazards regression was used to estimate hazard ratios with 95% confidence intervals. Time-to-event outcomes were calculated from completion of primary treatment to the first documented recurrence, metastasis, or death from any cause; patients without events were censored at last follow-up. Statistical significance was defined as p < 0.05. Analyses were performed using IBM SPSS Statistics version 23.0.

Results

Between 2012 and 2014, a total of 164 Thai cervical cancer patients were enrolled in this study. Fourteen patients were excluded—13 due to loss to follow-up and 1 who declined standard treatment—all of whom were from the HPV DNA–positive group. The remaining 150 patients, all of whom completed primary treatment and for whom complete medical records were available, were included in the final analysis. This cohort consisted of 139 patients (92.7%) with HPV DNA–positive cervical cancer and 11 patients (7.3%) with HPV DNA–negative cervical cancer (Figure 1).

Figure 1 Enrollment, follow-up, and analysis of participants.

Significant differences were observed in adenosquamous histologic subtype, staging, and primary treatment modality between the two groups (Table 1). Histological subtype analysis revealed a significantly higher proportion of adenosquamous carcinoma in the HPV DNA–negative group (p = 0.027). Overall stage distribution differed between groups (p = 0.026). HPV DNA–negative tumors were more frequently diagnosed at stage I, whereas HPV DNA–positive tumors were more often stage II–III; however, comparisons at individual stage levels did not reach statistical significance. Treatment patterns differed significantly: surgery was more common in the HPV DNA–negative group (36.4% vs. 10.8%), whereas definitive chemoradiotherapy was more frequently administered to HPV DNA–positive cases (84.9% vs. 54.5%), with statistically significant differences (p = 0.014 and 0.010, respectively). These differences in stage and treatment between groups represent potential confounders that should be considered when interpreting survival comparisons.

Table 1 Baseline Characteristics of the Participants (n=150)

No significant differences were found in patient age, tumor size, presence of lymph node metastasis, or distribution of squamous or adenocarcinoma histological subtypes (Table 1). The mean ages of patients with HPV DNA–positive and HPV DNA–negative cervical cancer were 54.4 and 50.9 years, respectively (p = 0.342). Mean tumor size did not differ significantly between groups (p = 0.463). The rates of lymph node metastasis were similar between the two groups (HPV DNA–positive: 27.3% vs. HPV DNA–negative: 18.2%; p = 0.511). Squamous cell carcinoma was the most common histologic subtype in both groups, with a higher prevalence in the HPV DNA–positive group (72.3%) compared to the HPV DNA–negative group (54.6%), although the difference was not statistically significant (p = 0.111). Similarly, adenocarcinoma was more frequent in HPV DNA–negative cases (27.3%) than in HPV DNA–positive cases (19.1%), but this difference was also not statistically significant (p = 0.574).

Treatment responses after completion of primary treatment and recurrence rates were comparable between the two groups (Table 1). Specifically, complete response was achieved in 83.5% of HPV DNA–positive cases and 90.9% of HPV DNA–negative ones. Moreover, disease recurrence within the first 5 years occurred in 36.0% of HPV DNA–positive cases and 36.4% of HPV DNA–negative cases (p = 0.608). Persistent disease and post-treatment progression were observed in 8.6% and 7.9% of HPV DNA–positive cases, and 0.0% and 9.1% of HPV DNA–negative cases, respectively (p = 0.311 and 0.888).

The data cut-off date was November 2024. The median follow-up period was 9.8 years. At the 5-year mark, 54 of 150 patients experienced recurrence, metastasis, or death; by 10 years, this number increased to 71 patients. The estimated 5-year PFS rate for the entire cohort was 62.2% (95% CI: 0.54–0.70), and the 10-year PFS rate was 49.7% (95% CI: 0.41–0.58) (Table 2).

Table 2 Summary of Kaplan–Meier Survival Estimates

Group-specific survival data are summarized in Table 2. Within 5 years, events occurred in 36.0% (50/139) of patients in the HPV DNA–positive group and 36.4% (4/11) in the HPV DNA–negative group; by 10 years, the respective event rates were 47.5% (66/139) and 36.4% (4/11). The 5-year PFS rates were 62.0% in the HPV DNA–positive group and 70.0% in the HPV DNA–negative group, with no statistically significant difference between groups (HR 1.06; 95% CI: 0.38–2.93; p = 0.916) (Figure 2A). Similarly, 10-year PFS rates of 49.6% and 52.5%, respectively, did not differ significantly (HR 1.42; 95% CI: 0.52–3.88; p = 0.501) (Figure 2B). The wide confidence intervals across both time points reflect the limited statistical power attributable to the small HPV DNA–negative subgroup, and the absence of statistical significance should not be interpreted as evidence of equivalent survival outcomes.

Figure 2 Kaplan–Meier estimates of progression-free survival (PFS) in the two cohorts. (A) Five-year PFS. (B) Ten-year PFS.

Overall survival trends were similar to those observed for PFS. Within 5 years, 30.9% (43/139) of HPV DNA–positive cases and 36.4% (4/11) of HPV DNA–negative cases experienced death from any cause. By 10 years, this increased to 52 (37.4%) in the former group but remained unchanged at 4 (36.4%) in the latter group. The 5-year OS rate was 66.7% in the HPV DNA–positive group and 63.6% in the HPV DNA–negative group (HR 0.93; 95% CI: 0.33–2.58; p = 0.885) (Figure 3A). At 10 years, OS was 56.8% and 63.6%, respectively (HR 1.14; 95% CI: 0.41–3.16; p = 0.799) (Figure 3B). These findings indicate no statistically significant differences in long-term OS between groups; however, given the small HPV DNA–negative subgroup and wide confidence intervals, the possibility of a clinically meaningful difference cannot be excluded.

Figure 3 Kaplan–Meier estimates of overall survival (OS) in the two cohorts. (A) Five-year OS. (B) Ten-year OS.

As a secondary outcome, the incidence of HPV DNA–negative cervical cancer at Chulabhorn Hospital during the study period was 7.3% (11 of 150 cases). Additionally, baseline characteristics, treatment responses, and oncological outcomes for each of the patients with HPV DNA–negative cervical cancer are detailed in Table 3. This table provides a case-by-case comparison of clinical presentation, histopathology, and treatment outcomes, highlighting the heterogeneity of HPV DNA–negative cervical cancer cases and the variability in their prognosis.

Table 3 Characteristics of Patients with HPV DNA–Negative Cervical Cancer

Discussion

This study evaluated long-term oncologic outcomes of HPV DNA–negative cervical cancer in a Thai tertiary-care cohort. HPV DNA–negative tumors accounted for 7.3% of cases, consistent with international estimates (approximately 5%–18%).^{8–10,14–17} Although HPV-independent cervical cancer, as defined in WHO classifications incorporating both HPV and p16 testing, has been associated with more aggressive behavior and inferior survival in several cohorts, our HPV DNA–defined negative group did not demonstrate significantly different PFS or OS compared with HPV DNA–positive patients. These findings should be interpreted cautiously, as the small number of HPV DNA–negative cases (n=11) resulted in wide confidence intervals and substantially limited statistical power; consequently, the absence of statistical significance does not constitute evidence of equivalent outcomes.

Clinicopathologically, HPV DNA–negative tumors in our cohort were more frequently associated with adenosquamous histology and earlier-stage disease. The higher proportion of stage I diagnoses in the HPV DNA–negative group contrasts with prior reports linking HPV-independent tumors to advanced-stage presentation. This difference in stage distribution may partly explain the absence of survival disadvantage in our study, as stage at diagnosis remains the strongest determinant of outcome in cervical cancer. Additionally, the higher proportion of HPV DNA–negative patients who received primary surgery, which may confer better local disease control in early-stage tumors, may have further contributed to the observed outcomes. These treatment and stage differences represent potential confounders that limit the interpretability of unadjusted survival comparisons between groups.

Previous international studies have reported substantially worse survival among HPV-independent tumors. Stolnicu et al observed a marked reduction in recurrence-free survival, while Kaliff et al and Nicolás et al reported significantly lower 5-year OS.^7–9 Moreover, previous studies by da Mata et al, Li et al, and Kugelman et al suggested that positivity for HPV DNA was associated with a good prognosis.^17–19 The discrepancy between those findings and ours may reflect several factors. First, our study relied solely on HPV DNA genotyping without p16 immunohistochemistry, which is recommended by WHO to support classification of HPV-independent tumors; misclassification of true HPV-associated tumors as HPV DNA–negative cannot be excluded. Second, the favorable stage distribution and higher surgical rates in our HPV DNA–negative group may partially explain the lack of a survival difference. Third, biological heterogeneity of HPV DNA–negative tumors—which may encompass both true HPV-independent cancers and false-negative cases with low or undetectable viral load—could contribute to the variable outcomes observed. The “hit-and-run” hypothesis further suggests that HPV-driven carcinogenesis may persist even after viral DNA becomes undetectable.^4,20

The possibility of false-negative HPV results should also be considered, particularly in cases with low viral load, non-targeted genotypes, or degraded DNA. False-negative HPV DNA results could lead to misclassification of truly HPV-associated tumors into the HPV DNA–negative category, potentially attenuating any apparent survival difference. The absence of p16 immunohistochemistry, which is recommended in the WHO classification to differentiate HPV-associated from HPV-independent tumors, represents a key limitation of this study. Future studies should incorporate combined HPV DNA testing and p16 immunohistochemistry to enable definitive WHO-based classification.

This study benefits from an extended median follow-up of nearly 10 years, standardized staging using FIGO 2018 criteria, and consistent institutional treatment protocols. However, several limitations must be acknowledged. Most importantly, the small number of HPV DNA–negative cases (n=11) substantially limits statistical power and may obscure true survival differences between groups; wide confidence intervals further underscore this uncertainty. HPV status was determined solely by HPV DNA testing without p16 immunohistochemical confirmation, precluding definitive WHO-based classification and introducing potential misclassification bias. The retrospective single-center design may introduce selection bias and limits generalizability, particularly to broader Southeast Asian populations. Notable differences in stage distribution and primary treatment modality between groups represent additional potential confounders. Future multicenter prospective studies with larger cohorts and standardized molecular classification—incorporating both HPV DNA testing and p16 immunohistochemistry—are essential to clarify the true prognostic significance of HPV-independent cervical cancer.

Taken together, our findings suggest that HPV DNA–negative cervical cancer may represent a heterogeneous entity with variable biological behavior. In this Thai cohort, survival outcomes did not demonstrate the adverse prognosis reported elsewhere; however, given the exploratory nature of these findings and the substantial methodological limitations, this study should be regarded as hypothesis-generating rather than conclusive. The importance of regional data and the need for standardized molecular classification are underscored.

Conclusion

In this exploratory retrospective analysis of a Thai tertiary cohort, HPV DNA–negative cervical cancer accounted for 7.3% of cases. No statistically significant differences in 5-year or 10-year PFS or OS were observed between HPV DNA–negative and HPV DNA–positive patients; however, these results must be interpreted with caution given the small HPV DNA–negative sample size (n=11), which substantially limits statistical power and precludes definitive conclusions regarding equivalence of outcomes. Differences in stage distribution and treatment modality between groups represent additional potential confounders. The absence of p16 immunohistochemistry further constrains interpretation within the WHO classification framework. Larger prospective multicenter studies incorporating standardized HPV DNA and p16 testing are required to clarify the prognostic implications of HPV DNA–negative, and specifically WHO-defined HPV-independent, cervical cancer.