Characteristics of solid versus hematologic malignancies in people living with human immunodeficiency virus

Abstract

Purpose

People living with HIV (PLWH) face increased risks of both AIDS-defining and non-AIDS-defining malignancies. We characterized the age-standardized cancer burden, compared clinical characteristics and survival outcomes between hematologic and solid malignancies in PLWH, and performed a secondary analysis stratifying malignancies by viral oncogenesis.

Methods

We conducted a retrospective cohort study of 166 PLWH diagnosed with cancer at a 407-bed tertiary academic medical center in the southern United States between 2019 and 2025. Age-standardized cancer rates proportions were calculated using the direct standardization method with the US 2000 Standard Population as the reference. Demographic, HIV-related, and clinical variables were compared between hematologic and solid malignancies using univariate and multivariate logistic regression. Cox proportional hazards regression and restricted mean survival time analysis were performed to identify predictors of overall survival. The analysis was repeated for virus-driven versus non-virus-driven malignancies.

Results

Among 3,640 HIV-positive patients actively receiving care at our institution, 166 (4.6%) were identified as having both HIV infection and a concurrent malignancy. The age-standardized cancer detection proportion was 888.1 per 100,000 patients (95% CI 753.0–1023.2). Hematologic malignancies occurred at a younger age than solid tumors (median 41 vs. 51 years, p = 0.002) and were associated with higher HIV viral loads, lower CD4 + counts, and poorer antiretroviral therapy adherence. High HIV viral load remained independently associated with hematologic malignancy in multivariate analysis (OR 4.839, 95% CI 1.495–17.58). Kaplan–Meier analysis demonstrated a trend toward lower overall survival among hematologic malignancy patients compared with solid tumor patients (log-rank p = 0.056), which reached significance after age adjustment in multivariate Cox proportional hazards regression (HR 3.00, 95% CI 1.35–6.66, p = 0.007). In the virus-driven analysis, both high HIV viral load and low CD4 + T-cell count were independently associated with virus-driven malignancy, with CD4 + T-cell count demonstrating a particularly strong association (OR 5.129, 95% CI 1.709–17.153, p = 0.005).

Conclusion

Immune dysfunction remains strongly associated with hematologic malignancies in PLWH, whereas solid tumors are increasingly associated with age- and lifestyle-related factors. A secondary analysis stratifying malignancies by viral oncogenesis identified consistent associations between immune dysfunction and virus-driven malignancy, further supporting the role of impaired immune surveillance in HIV-associated cancer development.

Clinical trial number

Not applicable.

Introduction

Since the discovery of human immunodeficiency virus (HIV) in 1984, Kaposi sarcoma (KS), non-Hodgkin lymphoma (NHL), and cervical cancer have been the most frequently reported malignancies among people living with HIV (PLWH) [1–3]. These tumors are classified as acquired immunodeficiency syndrome (AIDS)–defining cancers (ADCs) [2–5]. Despite the introduction of antiretroviral therapy (ART) in 1996, cancer remains a leading cause of death among PLWH, accounting for 10–20% of HIV-related deaths [2, 3, 6]. Hematologic malignancies, including lymphomas and leukemias arising from blood-forming cells, occur at substantially higher rates among PLWH than in the general population, with non-Hodgkin lymphoma occurring 11 times the expected rate and Hodgkin lymphoma occurring at approximately 7 to 10 times the expected rates respectively [7, 8]. Solid tumors, which include non-hematologic cancers arising from organ or tissue cells such as lung, prostate, anal, and colorectal cancers, are overall approximately 1.7 times more common among PLWH than the general population and represent a growing proportion of HIV-associated malignancies as ART extends survival, allowing aging-related and lifestyle-related risk factors to accumulate over time [2, 7].

Coinciding with this shift, cancer incidence patterns have shifted, with non–AIDS-defining cancers (NADCs) such as lung cancer, classic Hodgkin lymphoma, anal cancer, and hepatocellular carcinoma (HCC) now exceeding ADCs as the predominant malignancies [1, 2, 9–11]. Nationally among PLWH, ADC burden is projected to decline from 2,730 cases in 2010 to 720 cases by 2030, while solid tumor NADCs are projected to increase, with prostate and lung cancers emerging as the most common malignancies in this population [12]. The southern United States bears a particularly substantial HIV-associated cancer burden, with anal cancer incidence among PLWH in the South more than fourfold higher than in other US regions [13]. Despite this disproportionate regional burden, contemporary data specifically characterizing the distribution and clinical predictors of solid versus hematologic malignancies among PLWH in the southern United States remain limited.

The risk factors driving hematologic and solid malignancies among PLWH differ substantially. Hematologic malignancies are primarily associated with immune dysfunction, with low CD4 + T-cell count, high HIV viral load, Epstein-Barr virus (EBV), and poor ART adherence identified as key predictors of lymphoma development in this population [8, 14]. In contrast, solid tumors among PLWH are increasingly attributable to traditional lifestyle and behavioral risk factors including alcohol use, metabolic comorbidities such as hypertension, and older age [2, 15]. Notably, CD4 + T-cell count and viral load predict infection-related solid tumors but show little association with infection-unrelated solid tumors, suggesting that immune status and behavioral risk factors make distinct contributions to malignancy development in this population [14].

Prior studies have examined cancer risk in PLWH using either AIDS-defining status or viral oncogenesis as classification frameworks [7, 14, 16]. The present study extends prior work by simultaneously applying both a hematologic versus solid tumor comparison and a virus-driven versus non-virus-driven classification within a contemporary southern U.S. cohort. ART adherence, HIV viral load, and CD4 + T-cell count were incorporated as predictors across both analytical frameworks. This retrospective study aims not only to describe the distribution, age-standardized cancer rate proportion, and clinical profiles of malignancies among PLWH, but also to identify predictors of malignancy type and examine survival outcomes across both classification frameworks.

Methods

We conducted a retrospective cohort study of patients diagnosed with both HIV and cancer between November 1, 2019, and April 30, 2025, at Ochsner LSU Health Shreveport – Academic Medical Center, a 407-bed nonprofit tertiary academic medical center and the only Level 1 trauma center in northwest Louisiana, serving patients from Louisiana, Arkansas, Texas, and Mississippi. The institution includes the Feist-Weiller Cancer Center providing comprehensive oncology services and a dedicated Viral Disease Clinic offering inpatient and outpatient HIV care to approximately 1200 active patients through the Division of Infectious Diseases. Higly active antiretroviral therapy (HAART) regimen has been available at this institution since 1996 and has been accessible to all HIV-positive patients regardless of insurance status. This site was selected as it represents the primary tertiary referral center for HIV and oncology care in the region and it is one of the highest HIV burden areas in the United States. Eligible patients were identified using ICD-10 diagnostic codes for HIV infection (B20) and concurrent malignancy through Epic Slicer Dicer, followed by manual chart review to confirm eligibility and extract clinical data. All patients with established HIV infection prior to cancer diagnosis who received care at this institution during the study period were included, representing a consecutive census sample rather than a selected subset. A formal sample size calculation was not performed as this was a retrospective cohort study including all eligible patients identified during the study period. Patients with a history of cancer prior to HIV diagnosis were excluded. Demographic, clinical, and HIV-related data were collected from electronic medical records.

Hematologic malignancies included lymphoma and leukemia, while solid tumors encompassed all non-hematologic cancers. Patients with CD4 + T-cell counts ≥ 200 cells/mm³ and HIV viral load < 200 copies/mL were considered to have well-controlled HIV infection. The most recent CD4 + T-cell count and HIV viral load obtained within 12 months preceding cancer diagnosis were used in all analyses. ART adherence was operationalized as a binary variable based on clinician documentation in the electronic medical record, with patients classified as adherent or nonadherent based on the most recent clinical assessment prior to cancer diagnosis.

Malignancies were additionally classified as virus-driven or non-virus-driven based on established oncogenic viral etiologies. Virus-driven malignancies included those with well-established causative viral associations: HHV-8 for Kaposi sarcoma, EBV for lymphoid malignancies including non-Hodgkin lymphoma, Hodgkin lymphoma, Burkitt lymphoma, CNS lymphoma, and plasmablastic lymphoma, HPV for anogenital and cervical cancers, and HBV or HCV for hepatocellular carcinoma. For DLBCL, individual pathology reports were reviewed for EBER in situ hybridization results. Cases without EBER documentation were classified as virus-driven only when supported by additional clinical evidence of viral oncogenesis, including concurrent virus-associated malignancies, documented oncogenic viral co-infections, or clinical suspicion documented by the treating team in the context of severe immunocompromise. Follicular lymphoma, small lymphocytic lymphoma, nodal marginal zone lymphoma, and B-cell lymphoblastic lymphoma were classified as non-virus-driven given infrequent EBV association in the literature. For squamous cell carcinomas at sites with variable HPV attribution rates, including oropharyngeal, oral, and unspecified sites, virus-driven classification required explicit HPV documentation in the medical record or pathology report.

Numerical variables were summarized using descriptive statistics. Data normality was assessed using the Kolmogorov–Smirnov test. Normally distributed variables were analyzed using Welch’s t-test, and categorical variables were analyzed using Fisher’s exact test. Univariate logistic regression identified factors associated with hematologic and solid malignancies, and variables with p < 0.10 were entered into a multivariate logistic regression model to identify independent predictors of hematologic malignancy. A separate multivariate logistic regression model was constructed using virus-driven versus non-virus-driven malignancy as the outcome, with HIV viral load, CD4 + T-cell count, and ART adherence as predictors based on their established biological relevance to viral oncogenesis. Model performance for both multivariate models was assessed using receiver operating characteristic analysis and the Hosmer–Lemeshow goodness-of-fit test, and the resulting curves were compared. Univariate and multivariate Cox proportional hazards regression was additionally performed to identify predictors of overall survival, with variables significant at p < 0.10 in univariate analysis entered into the multivariate model. The proportional hazards assumption was verified using the cox.zph function. Age-standardized cancer detection proportions were calculated using the direct standardization method with the US 2000 Standard Population as the reference, expressed per 100,000 patients. Rates were calculated overall and separately for hematologic, solid tumor, virus-driven, and non-virus-driven malignancies. Kaplan–Meier analysis with log-rank testing compared overall survival between malignancy groups. As median survival could not be estimated for all groups due to incomplete events, the restricted mean survival time was calculated at tau = 60 months using the survRM2 package in R. Statistical significance was defined as p < 0.05. Analyses were performed in GraphPad Prism (version 10.6) and R Studio (version 2025.09.2 + 418), with the survival and survRM2 packages used for Cox regression and restricted mean survival time analyses.

Results

Among 3,640 HIV-positive patients received care at our institution during the study period, 166 (4.6%) were identified as having both HIV infection and a concurrent malignancy. The age-standardized cancer detection proportion was 888.1 per 100,000 patients (95% CI 753.0–1023.2). The age-standardized detection proportion for hematologic malignancies was 233.4 per 100,000 (95% CI 162.0–304.9) and for solid tumors was 654.7 per 100,000 (95% CI 539.9–769.5). The age-standardized detection proportion for virus-driven malignancies was 522.1 per 100,000 (95% CI 406.3–638.0) and for non-virus-driven malignancies was 366.0 per 100,000 (95% CI 289.5–442.5).

The cohort had a median age of 50 years (range, 14–81); 80.1% (133/166) were male. Of the cohort, 126 patients (75.9%) were African American and 39 (23.5%) were Caucasian. Among 98 patients who reported sexual orientation, 52 (53.1%) identified as homosexual or bisexual and 46 (46.9%) as heterosexual. Smoking history was documented in 113 cases (68.1%), alcohol use in 60 (36.6%), and illicit drug use in 80 (50%). Common comorbidities included hypertension (90, 54.2%), asthma or COPD (31, 18.7%), and chronic kidney disease or end-stage renal disease (29, 17.5%) (Table 1).

Table 1.

Baseline demographic and clinical characteristics of people living with HIV stratified by tumor type

Demographic and clinical characteristics of 166 people living with HIV diagnosed with malignancy, stratified by tumor type (solid vs. hematologic). Continuous variables are presented as median (range) and categorical variables as number (percentage). P values were calculated using Welch’s t-test for continuous variables and Fisher’s exact test for categorical variables. Statistically significant values (P < 0.05) are bolded. Note: CKD, chronic kidney disease; ESRD, end-stage renal disease; ART, antiretroviral therapy; CD4 Count (cells/µL); HIV Viral Load (copies/mL)

The median interval from HIV to cancer diagnosis was 10 years (range, 0–37); 15 patients (10.7%) were diagnosed concurrently. The most frequent malignancies included non-Hodgkin lymphoma (18%), anal cancer (12%), prostate cancer (10.7%), and Kaposi sarcoma (9.5%) (Fig. 1Panel A). AIDS-defining cancers accounted for 24.7% of all malignancies. Before cancer diagnosis, 79 of 115 patients (68.7%) had CD4 + T-cell counts ≥ 200 cells/mm³, and 62 of 109 (56.9%) had HIV viral loads < 200 copies/mL. Among 141 patients, 67 (47.5%) were adherent to ART, 59 (41.8%) were nonadherent, and 15 (10.7%) were diagnosed concurrently.

Fig. 1.

Clinical characteristics, survival outcomes, and factors associated with hematologic, solid, virus-driven, and non-virus-driven malignancies among people living with HIV. (A) Distribution of malignancy types diagnosed among 166 people living with HIV. Bars represent the number of cases within each cancer category, ordered from most to least common. Non–AIDS-defining cancers comprised the majority of cases, whereas AIDS-defining cancers such as Kaposi sarcoma and non-Hodgkin lymphoma were less frequent. Red bars indicate AIDS-defining malignancies and blue bars indicate non-AIDS-defining malignancies. (B) Kaplan–Meier overall survival curves comparing hematologic and solid malignancies. Shaded regions denote 95% confidence intervals. The log-rank test showed no statistically significant difference between groups (p = 0.056). (C) Forest plot of univariate logistic regression evaluating risk factors associated with hematologic malignancy. Odds ratios with 95% confidence intervals are shown for each variable. The dashed vertical line indicates the null value (odds ratio = 1). (D) Forest plot of univariate logistic regression evaluating risk factors associated with solid tumor malignancy. Odds ratios with 95% confidence intervals are displayed. The dashed vertical line represents the null value (odds ratio = 1). (E) Forest plot of univariate logistic regression evaluating immune-related risk factors associated with virus-driven malignancy. Odds ratios with 95% confidence intervals are shown for high HIV viral load (OR 3.146, p = 0.004), low CD4 + T-cell count (OR 5.571, p < 0.001), and ART nonadherence (OR 2.468, p = 0.014). The dashed vertical line indicates the null value (odds ratio = 1). All three variables were significantly associated with virus-driven malignancy. (F) Receiver operating characteristic curves comparing multivariate logistic regression model performance. The blue curve represents the hematologic versus solid malignancy model (AUC = 0.708, 95% CI 0.596–0.821). The red curve represents the virus-driven versus non-virus-driven malignancy model (AUC = 0.717, 95% CI 0.613–0.820). The dashed diagonal line represents the reference for no discriminatory ability (AUC = 0.50). Note: ART, antiretroviral therapy

At one-year follow-up, 151 patients (90.96%) were alive and 15 (9.04%) had died. Solid tumors were identified in 125 patients (75.3%) and hematologic malignancies in 41 (24.7%). Patients with hematologic malignancies demonstrated a trend toward lower overall survival compared with those with solid tumors (log-rank p = 0.056) (Fig. 1Panel B). Patients with hematologic malignancies were younger at diagnosis and had higher HIV viral loads (≥ 200 copies/mL in 70% vs. 32.9%, p = 0.001), lower CD4 + counts (< 200 cells/mm³ in 48.4% vs. 25%, p = 0.017), and lower ART adherence (30% vs. 54.5%, p = 0.04). Alcohol use (p = 0.02) and hypertension (p = 0.01) were more common among patients with solid tumors.

In univariate Cox proportional hazards regression, malignancy type demonstrated a trend toward association with overall survival (solid vs. hematologic HR 2.031, 95% CI 0.967–4.265, p = 0.061) and older age was significantly associated with worse survival (HR 1.032, 95% CI 1.003–1.062, p = 0.030). HIV viral load, CD4 + T-cell count, ART adherence, and virus-driven classification were not significantly associated with overall survival (p = 0.695, 0.111, 0.719, and 0.888 respectively). In multivariate analysis adjusting for age, solid tumor patients demonstrated significantly better overall survival compared with hematologic malignancy patients (HR 3.00, 95% CI 1.35–6.66, p = 0.007), and older age remained independently associated with worse survival (HR 1.046 per year, 95% CI 1.013–1.079, p = 0.005). The proportional hazards assumption was confirmed for all variables (global p = 0.81). The restricted mean survival time at 60 months was 47.99 months for hematologic malignancy patients versus 52.81 months for solid tumor patients (difference − 4.82 months, p = 0.221). The restricted mean survival time did not differ meaningfully between virus-driven and non-virus-driven malignancies (52.06 vs. 51.01 months, difference 1.05 months, p = 0.736).

Univariate regression identified high HIV viral load, low CD4 + count, and ART nonadherence as factors associated with hematologic malignancy, while alcohol use and hypertension were associated with solid tumors (Fig. 1Panel C and Panel D). In multivariate analysis, high HIV viral load remained independently associated with hematologic malignancy (OR 4.839, 95% CI 1.495–17.58). CD4 + count ≥ 200 cells/mm³ (OR 1.734, 95% CI 0.571–5.307) and ART adherence (OR 0.6715, 95% CI 0.1761–2.336) were not statistically significant. Model performance was acceptable (AUC 0.708, 95% CI 0.596–0.821, p < 0.001), with good calibration (Hosmer–Lemeshow p = 0.174) and a classification accuracy of 70.1%.

In a secondary analysis, cancers were classified as virus-driven (n = 78, 47.0%) or non-virus-driven (n = 88, 53.0%). In univariate analysis, high HIV viral load (OR 3.146, 95% CI 1.445–7.042, p = 0.004, n = 109), low CD4 + T-cell count (OR 5.571, 95% CI 2.376–14.099, p < 0.001, n = 116), and ART nonadherence (OR 2.468, 95% CI 1.213–5.118, p = 0.014, n = 126) were each significantly associated with virus-driven malignancy (Fig. 1Panel E). High HIV viral load (OR 3.299, 95% CI 1.111–10.463, p = 0.035) and low CD4 + T-cell count (OR 5.129, 95% CI 1.709–17.153, p = 0.005) remained independently associated with virus-driven malignancy in multivariate analysis, while ART adherence did not reach significance (OR 0.598, 95% CI 0.167–1.922, p = 0.405). The virus-driven model demonstrated comparable discrimination to the primary model (AUC 0.717, 95% CI 0.613–0.820 versus AUC 0.708, 95% CI 0.596–0.821; Fig. 1Panel F) with excellent calibration (Hosmer–Lemeshow p = 0.723).

Discussion

This study characterizes the cancer burden and clinical predictors of malignancy type among PLWH at a contemporary southern U.S. tertiary care center. The observed age-standardized cancer detection proportion is consistent with published estimates of approximately twice the cancer incidence among PLWH compared with the general population, validating the substantial oncologic burden in this population despite widespread ART availability [7]. Two distinct oncologic patterns emerged, with hematologic malignancies driven primarily by immune dysfunction and solid tumors increasingly associated with aging and lifestyle-related risk factors, and malignancy type and age were identified as the only independent predictors of overall survival after age adjustment.

The younger age at cancer diagnosis among PLWH compared with the general population is well established and was confirmed in our cohort, with hematologic malignancy patients diagnosed at a significantly younger age than solid tumor patients [9, 10, 17, 18]. As ART has extended survival, aging has become an increasingly important driver of cancer development among PLWH [3, 18]. Consistent with this shift, alcohol use and hypertension were more prevalent among solid tumor patients in our cohort, supporting the growing contribution of behavioral and metabolic comorbidities to solid tumor development in this population. A high prevalence of smoking was also documented in our cohort, which alongside alcohol use likely contributes to the growing burden of non-AIDS-defining cancers among PLWH [2, 15].

Viral co-infections continue to play an important role in HIV-associated oncogenesis. Epstein–Barr virus, human herpesvirus 8, and human papillomavirus are implicated in the development of non-Hodgkin lymphoma, Kaposi sarcoma, and cervical cancer, respectively [10]. PLWH also have higher rates of hepatitis B and C co-infection, contributing to an increased risk of hepatocellular carcinoma [2, 19]. In our cohort, all patients with hepatocellular carcinoma had underlying hepatitis B or C infection and cirrhosis, highlighting the combined influence of chronic viral infection and hepatic injury on cancer development.

The secondary analysis stratifying malignancies by viral oncogenesis identified consistent associations between immune dysfunction and virus-driven malignancy, with CD4 + T-cell count demonstrating a particularly strong independent association. This finding is consistent with prior evidence that HIV-induced CD4 + T-cell depletion creates a permissive environment for oncogenic viral replication and malignant transformation, particularly for EBV and KSHV-driven malignancies [20, 21]. These results suggest that the association between immune dysfunction and hematologic malignancy observed in the primary analysis may be partly explained by virus-driven oncogenic pathways, underscoring the importance of early and sustained viral suppression in reducing virus-associated cancer risk among PLWH.

The identification of immune dysfunction as a key driver of virus-associated malignancies has important prevention implications. Virus-associated malignancies in PLWH may be particularly amenable to targeted interventions including early HIV diagnosis, sustained viral suppression, vaccination, and tailored cancer screening strategies. HPV vaccination has demonstrated effectiveness in reducing HPV-related cancers and is recommended for PLWH, yet vaccination rates remain suboptimal in this population [22, 23]. Similarly, HBV vaccination and tenofovir-containing ART regimens provide important protection against HBV-related hepatocellular carcinoma in PLWH [24].

Hematologic malignancies in our cohort developed significantly earlier after HIV diagnosis than solid tumors and were more frequently AIDS-defining, consistent with the established association between immune dysfunction and hematologic oncogenesis in PLWH. Correspondingly lower ART adherence, higher HIV viral loads, and lower CD4 + T-cell counts among hematologic malignancy patients further support the role of incomplete viral suppression and immune dysregulation in the development of these malignancies. Although causal relationships cannot be inferred from this retrospective analysis, these findings suggest that immune dysfunction remains an important driver of hematologic malignancies in the current treatment era despite widespread ART availability.

In contrast, solid tumors developed substantially later after HIV diagnosis and were predominantly non-AIDS-defining, reflecting the increasing contribution of traditional and lifestyle-related risk factors as ART extends survival in this population. The higher prevalence of alcohol use and hypertension among solid tumor patients supports the growing role of behavioral and metabolic comorbidities in solid tumor development among PLWH with longer life expectancy and improved HIV control. Chronic inflammation, oxidative stress, and endothelial dysfunction linked to hypertension and metabolic syndrome have been associated with increased cancer risk in aging populations, and our findings are consistent with prior work demonstrating that cardiovascular and metabolic comorbidities are highly prevalent among older virologically suppressed PLWH [17, 18, 25]. As PLWH age, cancer risk appears increasingly influenced by traditional risk factors rather than immunosuppression alone, underscoring the need for comprehensive HIV care models that integrate cardiovascular risk reduction, behavioral counseling, and age-appropriate cancer screening [11, 26, 27]. This pattern may also reflect survivorship bias, as individuals with durable ART adherence live long enough to manifest age- and lifestyle-related malignancies.

Survival analysis demonstrated a trend toward worse overall survival among hematologic malignancy patients that reached significance after adjustment for age, with solid tumor patients demonstrating substantially better survival outcomes in the age-adjusted Cox proportional hazards model. The attenuation of the unadjusted survival difference after age adjustment suggests that the younger age of hematologic malignancy patients may have partially obscured the true survival disadvantage associated with malignancy type. The restricted mean survival time analysis similarly showed a numerical difference favoring solid tumor patients, though this did not reach statistical significance, likely reflecting the limited number of events in this cohort. The observed survival pattern is consistent with broader evidence suggesting that hematologic malignancies are associated with worse outcomes in immunocompromised populations, potentially reflecting the combined effects of poorer HIV control, greater immune suppression, and malignancy-specific factors. The mechanisms underlying HIV-related oncogenesis include chronic inflammation, proto-oncogene activation, impaired immune surveillance, and inhibition of tumor suppressor genes [28]. HIV proteins such as p17 may further promote lymphoid proliferation and lymphoma growth [10].

The convergence of findings across the classification frameworks reinforces a consistent biological narrative. Classical markers of HIV-related immune dysfunction, including high viral load, low CD4 + T-cell count, and poor ART adherence, were significantly associated with both hematologic and virus-driven malignancies, suggesting that these frameworks capture the same underlying phenomenon of immune dysregulation from different classificatory perspectives. Notably, the virus-driven classification framework identified a particularly strong association with CD4 + T-cell count compared with the hematologic versus solid tumor framework, suggesting that viral oncogenesis may offer complementary insight into the immunological substrate of HIV-associated oncogenesis. This consistency strengthens the argument that immune suppression remains a central driver of HIV-associated cancer risk regardless of how malignancies are categorized [20]. From a clinical standpoint, these findings highlight the importance of early and routine screening for oncogenic viral co-infections, including EBV, KSHV, HPV, and hepatitis B and C, in immunocompromised PLWH, as earlier detection may contribute to improved outcomes through timely intervention, targeted vaccination, and sustained viral suppression [22, 24]. Notably, overall survival did not differ significantly between virus-driven and non-virus-driven malignancies, suggesting that tissue of origin rather than viral etiology may be the primary determinant of survival differences in this cohort.

The demographic composition of this cohort warrants consideration when interpreting these findings. The predominance of male participants reflects the local epidemiology of HIV in the southern United States, where African American individuals bear a greater burden of both HIV infection and cancer-related mortality [29–32]. In the context of HIV, African American men who have sex with men have been shown to have higher rates of virus-related malignancies including Kaposi sarcoma compared with White counterparts, potentially reflecting inequities in access to sustained or engaged in HIV care [33]. These findings underscore the particular importance of integrated HIV and cancer care models in the southern United States, where structural barriers to care may compound the preexisting elevated cancer burden among PLWH. Access to regular healthcare has been identified as a key predictor of cancer screening uptake among PLWH, highlighting the role of tertiary academic medical centers such as ours in providing comprehensive cancer surveillance to this vulnerable population [34]. The predominance of male participants limits the applicability of these findings to women living with HIV, who may have distinct cancer risk profiles and differential responses to ART. Sex-based differences in hematologic malignancy outcomes and immune function are well recognized and remain incompletely characterized in the context of HIV [35, 36]. Additionally, as a single-center cohort from the southern United States, these findings may not be generalizable to non-US settings or to populations in low- and middle-income countries where HIV epidemiology, ART access, and cancer screening infrastructure differ substantially [33].

Our study has several limitations. The retrospective, single-center design limits causal inference and may reduce generalizability, particularly to women living with HIV and to non-US or resource-limited settings where HIV epidemiology and access to care differ substantially. ART adherence was assessed through clinician documentation in the electronic medical record rather than objective measures such as pharmacy refill data, and may therefore be subject to misclassification or information bias. Missing laboratory and clinical data due to incomplete chart documentation may have introduced selection bias and reduced statistical power for certain analyses. The limited sample size and follow-up period further constrain the ability to detect significant differences in survival outcomes between malignancy groups. A formal sample size calculation was not performed as this was a retrospective cohort study including all eligible patients identified during the study period, and the limited number of events constrains the statistical power of the Cox proportional hazards survival analysis. Age-standardized cancer detection proportions were calculated using institutional data and reflect cancer detection among PLWH actively engaged in care at a tertiary academic medical center rather than true population-level incidence rates, which would require linkage with population-based cancer and HIV surveillance registries. Age-standardized cancer detection proportions were calculated using the number of HIV-positive patients seen at our institution as the denominator rather than true person-time at risk, and should therefore be interpreted as institutional detection proportions rather than population-based incidence rates. Unmeasured confounders including cancer stage, treatment variables, socioeconomic factors, and duration of viral suppression may also influence the observed associations. The comparison of hematologic versus solid tumors and virus-driven versus non-virus-driven malignancies, while clinically and biologically informative, represents a broad dichotomization that encompasses heterogeneous diseases with distinct etiopathogenetic mechanisms, and direct comparisons between these categories may not fully capture the complexity of HIV-associated oncogenesis. Future multicenter studies with larger and more diverse cohorts, standardized adherence measures, and longer follow-up are needed to further characterize cancer risk and outcomes in this population.

Conclusions

In summary, our findings describe two dominant oncologic patterns among people living with HIV: (1) hematologic malignancies associated with poorer HIV control, including lower ART adherence and higher HIV viral loads, and (2) solid tumors more frequently associated with older age and lifestyle- and metabolic-related risk factors such as alcohol use and hypertension. A secondary analysis stratifying malignancies by viral oncogenesis revealed consistent associations between immune dysfunction and virus-driven malignancy, suggesting that the classification frameworks capture the same underlying phenomenon of immune dysregulation from different classificatory perspectives. These findings underscore the importance of early HIV diagnosis, sustained viral suppression, targeted vaccination, and age-appropriate cancer screening in reducing the future cancer burden among PLWH. Comprehensive management addressing both immune restoration and modifiable comorbidities remains essential in this population.

Abbreviations

PLWH

People living with HIV

HIV

Human immunodeficiency virus

AIDS

Acquired immunodeficiency syndrome

KS

Kaposi sarcoma

NHL

Non-Hodgkin lymphoma

ADCs

AIDS-defining cancers

ART

Antiretroviral therapy

NADCs

Non-AIDS-defining cancers

HCC

Hepatocellular carcinoma

OR

Odds ratio

CI

Confidence interval

AUC

Area under the curve

ROC

Receiver operating characteristic

HHV

8-human herpesvirus 8

HPV

Human papillomavirus

EBV

Epstein-Barr virus

HBV

Hepatitis B virus

HCV

Hepatitis C virus

DLBCL

Diffuse large B cell lymphoma

EBER

Epstein-Barr encoding region in situ hybridization

FISH

Fluorescence in situ hybridization

KSHV

Kaposi sarcoma herpesvirus

IQR

Interquartile range

COPD

Chronic obstructive pulmonary disease

CKD

Chronic kidney disease

ESRD

End-stage renal disease

Author contributions

A.M.S. contributed to data collection, statistical analysis, data interpretation, drafting the manuscript, critical revision, and supervision. N.D. contributed to study conception and data collection. W.P.J. contributed to drafting the manuscript and critical revision. J.V. contributed to critical revision and supervision. A.E.M. contributed to study conception, data interpretation, supervision, revision, and drafting the manuscript. All authors reviewed and approved the final version of the manuscript.

Funding

The authors did not receive support from any organization for the submitted work.

Data availability

The datasets generated and/or analysed during the current study are not publicly available due to the sensitive nature of the data and the difficulty of effectively de-identifying the data due to a small, specific patient population.

Declarations

Ethical approval

The study protocol received approval from the LSU Health Shreveport Institutional Review Board (IRB STUDY00001921). Our study adhered to the Declaration of Helsinki. The Institutional Review Board of LSU Health Shreveport has agreed to waive consent of the participants because the research involved no direct patient contact and used existing, deidentified clinical data, the requirement for informed consent was waived.

Consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.