Cervical and Tumor-Associated Microbiome in Botswana Women With and Without HIV With Carcinoma of the Cervix Before and After Definitive Chemoradiation

Abstract

Background:

Cervical cancer remains a significant public health concern globally and particularly in sub-Saharan Africa, where high rates of HIV infection exacerbate cervical cancer incidence. Understanding the cervical microbiome and its role in cancer progression is essential, especially in regions where both cervical cancer incidence and HIV prevalence are high.

Purpose:

This study aimed to characterize the cervical microbiome in women living with HIV (WLWH) and HIV-negative women with squamous cell carcinoma of the cervix in Botswana, compare the microbiome between before and after chemoradiation therapy (CRT) in WLWH, and assess the prognostic value of specific microbial taxa for overall survival (OS) in WLWH.

Patients and Methods:

Cervical samples were collected from women with cervical cancer presenting to 1 hospital in 2018–2019. Patients’ clinical data, including HIV status, were recorded. Microbial composition was analyzed using 16S rRNA gene sequencing. Microbiome diversity and composition were evaluated using alpha and beta diversity metrics. Differential microbial abundance was analyzed using linear discriminant analysis effect size. The association between microbial taxa and OS was explored using Cox proportional hazards regression.

Results:

WLWH (n=42) had a significantly lower Pielou evenness index than HIV-negative women (n=11) (0.6 vs 0.7, p=0.02), suggesting a more imbalanced microbiome in WLWH. WLWH had higher levels of Parvimonas and members of the Corynebacteriaceae and Micrococcaceae families, suggesting a shift toward a more pathogenic microbiome. In WLWH, CRT did not significantly alter overall microbial diversity. However, Lactobacillus and Sutterella were enriched before treatment, reflecting a less pathogenic microbiome, while Ruminococcus and Phasocolarctobacterium and the families Caulobacterales and Flavobacteriia were enriched after treatment, reflecting microbial adaptations to the altered immune and treatment environment. Notably, higher levels of Flavobacteriia after CRT were independently associated with worse OS in WLWH.

Conclusion:

Microbiome profiles differ between WLWH and HIV-negative women with cervical cancer in Botswana. The microbiome may have prognostic significance. Future research is needed to better understand the significance of the microbiota in cervical cancer progression and treatment outcomes and the potential role of microbiome-targeted interventions.

INTRODUCTION

Cervical cancer is the leading cause of cancer-related death among women in Africa.^{1, 2} With over 660,000 new cases and more than 350,000 deaths globally in 2022 alone, the urgency to address this public health challenge cannot be overstated.³ Alarmingly, African women face a higher risk of cervical cancer than women in regions with better access to preventive healthcare screening.^{1, 4} Sub-Saharan Africa accounts for substantial proportions of cervical cancer cases and cervical cancer–related fatalities globally,^{5, 6} and projections suggest a significant increase in cervical cancer incidence in this region by 2030.⁷ Human papillomavirus (HPV) infection is recognized as a primary risk factor to cervical cancer, and thus there is a critical need for effective cervical cancer prevention strategies in regions, such as sub-Saharan Africa, where the prevalence of HPV and co-existing conditions like HIV is notably high.⁸

Despite strides made in cervical cancer screening programs, cervical cancer incidence remains alarmingly high, particularly among women living with HIV (WLWH).⁹ This paradox highlights the complex interplay between infectious diseases and cancer and underscores the urgent need for comprehensive approaches to disease prevention and management in sub-Saharan Africa, where the prevalence of HIV is high.

Research continues to shed light on the role of the cervical microbiome in cancer development and treatment outcomes. Our previous research among women in Botswana showed greater microbial diversity in women with cervical cancer than in those with cervical dysplasia, as well as distinctive differences in cervical microbiota composition between these groups.¹⁰ The study reported here built on that foundation by characterizing the cervical and tumor-associated microbiome in Botswana WLWH with squamous cell carcinoma of the cervix before and after treatment with chemoradiation and assessing whether specific microbial taxa predicted overall survival (OS). The highest incidence of cervical cancer exists in Sub-Saharan Africa¹¹, but there are significant knowledge gaps in understanding how the cervical microbiota affects treatment outcomes in this historically understudied and vulnerable population. To date, no published studies have specifically explored the cervical tumor microbiome and heterogeneity in WLWH in Botswana throughout their treatment.

PATIENTS AND METHODS

Participants and Clinical Data

Between July 2018 and February 2019, patients presenting at Princess Marina Hospital with newly diagnosed, biopsy-proven locally advanced, nonmetastatic cervical carcinoma (International Federation of Gynecology and Obstetrics [FIGO] staging system, 2009) were prospectively identified. Inclusion criteria were biopsy-proven stage IB2-IVA, nonmetastatic, cervical carcinoma. Exclusion criteria encompassed any history of noncervical primary cancer and ongoing pregnancy. Women who met these eligibility criteria were invited to participate in this study. For each participant, a comprehensive medical history, including a review of current medications, was meticulously obtained through interviews conducted by clinical providers or trained study staff. Demographic and clinicopathologic data were extracted from patient medical records. Tissue samples were collected at two timepoints: before the initiation of definitive chemoradiation therapy (CRT) and at the end of treatment. CRT involved external beam radiation therapy and brachytherapy with concurrent cisplatin for invasive cancer.

The study protocol, encompassing subject recruitment and tissue sampling, received approval from the Institutional Review Boards at the University of Botswana (UBR/RES/IRB/BIO/045), the University of Pennsylvania (830039), and The University of Texas MD Anderson Cancer Center (MDACC 2014–0543). Informed consent was mandatory for study participation and was documented in writing by patients.

Sample Collection and DNA Extraction

Cervical samples were collected using a matrix-designed quick-release Isohelix swab. Samples were collected by rubbing the Isohelix swab against the visible cervical tumor at the time of pelvic examination. Samples were immersed in 1 mL of phosphate-buffered saline within 1 hour of collection and stored at −80 °C. Bacterial genomic DNA extraction was carried out using a MO BIO PowerSoil DNA Isolation Kit (MO BIO Laboratories). Subsequently, the samples were shipped to the US for further processing, including DNA processing and sequencing.

16S rRNA Gene Sequencing and Sequence Data Processing

Sequencing and data processing were performed as we previously described.¹⁰ 16S rRNA gene sequencing of the cervical samples was performed at the Alkek Center for Metagenomics and Microbiome Research at Baylor College of Medicine (Houston, Texas) using methods adapted from those used for the Human Microbiome Project.¹² The 16S rDNA V4 region was amplified by PCR using primers that contained sequencing adapters and single-end barcodes, allowing the pooling and direct sequencing of PCR products. Amplicons were sequenced on the MiSeq platform (Illumina) using the 2×250-bp paired-end protocol, yielding paired-end reads that overlapped almost completely. The sequence reads were de-multiplexed, quality filtered, and subsequently merged using USEARCH version 7.0.1090 (4). 16S rRNA gene sequences were clustered into operational taxonomic units at a similarity cut-off value of 97% using the UPARSE algorithm.¹³ To generate taxonomies, we mapped operational taxonomic units to an optimized version of the SILVA rRNA database containing the 16S v4 region. A custom script was used to construct an operational taxonomic units table from the output files generated, as described above, for downstream analyses of alpha diversity, beta diversity, and phylogenetic trends. Principal coordinates analysis was conducted to ensure no batch effects within the sample sets.

Microbial Diversity Metrics and Composition

Alpha (within-sample) diversity was evaluated using Pielou’s evenness, Shannon diversity index, Simpson index, Simpson’s evenness, observed features, and Fisher’s alpha. Pielou’s evenness index assesses how evenly individuals are distributed among the species present.¹² The Shannon diversity index captures both the richness and evenness of taxa within a sample. The Simpson index emphasizes the relative abundance of species contributing to overall richness. Simpson’s evenness index further refines this by measuring how evenly individuals are distributed across species, giving more weight to dominant taxa.^{13, 14} Observed Features metric counts the number of species detected with at least one read.¹⁵.Lastly, Fisher’s alpha index reflects the relationship between species number and their abundance.¹⁶Beta (between-sample) diversity was determined assessed by calculating using the weighted UniFrac distances to generate coordinates for each sample. Principal coordinate analysis (PCoA) was used to visualize the distances between groups. Permutational multivariate analysis of variance (PERMANOVA) was used to assess differences is means between groups. The relative abundance of microbial taxa, classes, and genera was compared between samples from WLWH and HIV-negative women and between samples from WLWH obtained before CRT (baseline) and at the end of treatment. Differentially abundant bacterial genera were identified on the basis of case status using linear discriminant analysis effect size¹⁷ applying the 1-against-all strategy with a threshold of 4 on the logarithmic linear discriminant analysis score for discriminative features and an α of 0.05 for the Kruskal-Wallis test among classes. The observed differences were further examined using paired-analysis Student’s t test.

We classified the vaginal microbial communities of tumor samples into community state types (CST) using VALENCIA.¹⁸ Vaginal community state types (CST)CST describe groupings of the vaginal microbiome that fall into 5 categories based on the dominant bacterial taxa. There are 5 CST, with 4 influenced by specific lactobacillus organisms (I, II, III, V) and one comprised of a diverse anaerobe-rich flora (IV). CST group IV is broken down into three subcategories (IV-A, IV-B and IV-C) with distinctive composition¹⁸.

Statistical Analysis

Patient characteristics studied included HPV status, age, body mass index, smoking status, International Federation of Gynecology and Obstetrics (FIGO) stage, pathologic subtype, radiation therapy type, and concurrent chemotherapy status. These characteristics were compared between WLWH and HIV-negative women. Fisher’s exact test was used to analyze categorical variables, while Welch’s t-test was used to analyze continuous variables.

Patient and tumor characteristics were analyzed by univariate and multivariate Cox regression models for OS. Covariates with a p-value of ≤ 0.1 in the univariate analysis were included in the multivariate model. OS was defined as the time from the initiation of therapy until death caused by cervical cancer. Hazard ratios (HRs) and 95% confidence intervals (CIs) were estimated using Cox proportional hazards regression, and statistical significance was assessed using Wald tests.

RESULTS

We characterized the 16S rRNA cervical microbiome in 53 cervical cancer patients. Their clinicopathologic data are summarized in Table 1. Most patients were HIV positive and had squamous cell cancer. Forty-two of the 53 patients (79%) were WLWH. The 42 WLWH had a mean age of 44.4 years, and the 11 HIV-negative women had a mean age of 60.6 years (p < 0.001). The mean body mass index was 25.1 kg/m² in WLWH and 29.9 kg/m² in HIV-negative women (p = 0.03).

Table 1.

Patient Characteristics

Variable	HIV-positive, N=42 (%)	HIV-negative, N=11 (%)	p-value
HPV status			1
HPV positive	37 (88.1)	10 (90.9)
HPV negative	1 (2.4)	0 (0)
missing information	4 (9.5)	1 (9.1)
Age	44.4 ± 8.8	60.6 ± 12.1	0.001
BMI	25.1 ± 6.9	29.9 ± 5.9	0.03
Smoking status			0.19
no	40 (95.2)	9 (81.8)
yes	2 (4.8)	2 (18.2)
FIGO stage			0.05
I/II	26 (61.9)	8 (72.7)
III/IV	15 (35.7)	1 (9.1)
missing information	1 (2.4)	2 (18.2)
Pathology			0.62
Squamous cell carcinoma	39 (92.8)	10 (90.9)
Other	2 (4.8)	1 (9.1)
missing information	1 (2.4)	0 (0)
Radiation external beam radiation therapy (EBRT)			0.3
EBRT	4 (9.5)	2 (18.2)
EBRT and brachytherapy	37 (88.1)	8 (72.7)
missing information	1 (2.4)	1 (9.1)
Concurrent Chemotherapy			1
Cisplatin	34 (81.0)	9 (81.8)
None	5 (11.9)	1 (9.1)
missing information	3 (7.1)	1 (9.1)
Alpha diversity
Pielou’s evenness	0.6 ± 0.2	0.7 ± 0.2	0.02
Simpson’s evenness	0.1 ± 0.1	0.3 ± 0.2	0.06
Observed features	73.2 ± 43.9	51.3 ± 30.8	0.07
Fisher’s alpha	8.9 ± 6.5	7.4 ± 3.8	0.32
Shannon index	3.7 ± 1.2	4.0 ± 1.0	0.39
Simpson index	0.8 ± 0.2	0.9 ± 0.1	0.17

Cervical Microbiota by HIV Status

We initially examined the cervical tumor microbiota in relation to HIV status. The most abundant genera for WLWH and HIV-negative women are shown in Figure 1A. We next analyzed the distribution and change in of vaginal community state types (CSTs) among WLWH and HIV-negative patients. (Figure 1BAt baseline, non-Lactobacillus-dominant CST IV-C was the most prevalent CST detected in both WLWH (32 of 42; 76%) and HIV-negative women (10 of 11; 91%). CST IV-B, which is characterized by diverse non-Lactobacillus species, was present in 24% (10 of 42) of samples from WLWH but no samples from HIV-negative women. CST-I, which is characterized by dominance of Lactobacillus crispatus, was present in 9% (1 of 11) of samples from HIV-negative women but no samples from WLWH. CST II, IV-A, and V were not detected in any women. Among WLWH, the proportion of women with CST IV-B decreased between baseline and the end of treatment, and the proportions of women with CST III or CST IV-C increased.

FIG 1.

Relative abundance of microbiota in Botswana women with cervical cancer by HIV status. (A) Results from 16S rRNA gene sequencing. (B) Results from community state types (CST) profiling. Baseline, before treatment. EOT, end of treatment.

Next, we evaluated the alpha diversity of the microbiome. Diversity of the microbial composition can be evaluated based on the number of types of organisms (richness), or “evenness” in which the Pielou index which evaluates how evenly organisms are distributed in the populations. The “Simpson” index combines the richness and the evenness into a single metric. There were no statistical differences in species richness or Simpson diversity between WLWH and HIV-negative women (Figure 2A). However, the Pielou evenness index was lower in WLWH than in HIV-negative women (0.6 vs 0.7, p = 0.02) (Figure 2A). Next, we evaluated beta diversity which assesses how different the composition is between two groups. Globally, there was no difference in beta diversity between WLWH and HIV-negative women (weighted UniFrac: PERMANOVA R²=0.0243, p = 0.32) (Figure 2B). UsingLEfSe, we aimed to identify specific bacterial which may be differentially enriched within our patient cohort (p < 0.05, linear discriminant analysis score > 4). Notably, the analysis showed enrichment of the families Micrococcaceae and Corynebacteriaceae, the order Oceanospirillales, and the genus Corynebacterium in WLWH and enrichment of the family Gemellaceae and the genus Nesterenkonia in HIV-negative women (Figure 2C).

FIG 2.

Microbiome diversity in Botswana women with cervical cancer by HIV status. (A) Alpha diversity. *p < 0.02. (B) Beta diversity. (C) Linear discriminant analysis (LDA) effect size.

Changes in the Cervical Microbiome Between Baseline and the End of Treatment in WLWH

In the WLWH, alpha and beta diversity remained stable between baseline and the end of treatment in all patients (Figure 3A, C). Linear discriminant analysis effect size showed enrichment of the families Alcaligenaceae and Lactobacillaceae and the genus Sutterella before treatment and enrichment of the genus Ruminococcus and order Caulobacterales after treatment (Figure 3B).

FIG 3.

Microbiome diversity in Botswana women with cervical cancer and HIV infection before (baseline) and at the end of treatment (EOT) with chemoradiation.

Baseline and End-of-Treatment Predictors of OS for WLWH Undergoing Chemoradiation

The results of the univariate and multivariate survival analyses are presented in table 2. Six WLWH were excluded from this analysis due to non–cervical cancer–related deaths. In the univariate Cox proportional hazards regression model predicting OS at baseline, FIGO stage III/IV (versus I/II) was a risk factor for OS (HR 3.58, 95% CI 1.13–11.40, p = 0.03). In the univariate Cox proportional hazards regression model predicting OS at the end of treatment, Simpson’s evenness (HR 0.01, 95% CI 0.00–1.92, p = 0.08) and class Flavobacteriia (HR 2.60E+21, 95% CI 0.00 to 1.31E+46, p = 0.09) were risk factors for OS. At baseline, multivariate survival analyses again identified higher FIGO stage as an independent prognostic factor for OS (HR 5.30E+00, 95% CI 1.34E+00 to 2.09E+01, p = 0.02). At the end of treatment, multivariate survival analyses again identified Flavobacteriia as an independent prognostic factor for OS (HR 3.81E+60, 95% CI 1.62E+19 to 8.99E+101, p = 0.004) (Figure 4).

Table 2.

Univariate and Multivariate Survival Analyses

Variable	N	HR	Cl_Lower	Cl_Upper	p_value	N	HR	Cl_Lower	Cl_Upper	p_value
Age	36	1.03	0.97	1.10	0.28
BMI	34	1.01	0.93	1.10	0.86
FIGO stage (−III/IV vs. I/Il)	35	3.58	1.13	11.40	0.03	25	5.30E+00	1.34E+00	2.09E+01	0.02
Pathology (−SCC vs. other)	36	1.31	0.17	10.24	0.80
baseline
alpha diversify
pielou’s evenness	36	2.09	0.09	49.98	0.65
simpson’s evenness	36	3.40	0.05	215.81	0.56
observed features	36	1.00	0.99	1.02	0.84
fisher’s alpha	36	1.01	0.91	1.11	0.92
shannon index	36	1.08	0.68	1.71	0.75
simpson index	36	1.98	0.13	31.21	0.63
taxa
g_Sutterella	36	0.00	0.00	3.04E+93	0.56
g_Lactobacillus	36	0.00	0.00	27459.52	0.40
o_Caulobacterales	36	0.00	0.00	Inf	1.00
c_Flavobacteriia	36	0.00	0.00	Inf	1.00
g_Corynebacterium	36	322.53	0.00	9.24E+14	0.69
g_Halomonas	36	0.00	0.00	Inf	0.85
f_Micrococcaceae	36	3.22E+08	0.00	6.74E+22	0.24
g_Parvimonas	36	0.00	0.00	1.02E+09	0.25
EOT
alpha diversify
pielou’s evenness	26	0.15	0.01	2.28	0.17
simpson’s evenness	26	0.01	0.00	1.92	0.08	25	4.63E-04	3.82E-08	5.62E+00	0.11
observed features	26	1.00	0.99	1.01	0.76
fisher’s alpha	26	0.99	0.94	1.06	0.86
shannon index	26	0.84	0.52	1.37	0.49
simpson index	26	0.17	0.01	4.90	0.30
taxa
g_Sutterella	26	3.65E+47	0.00	4.09E+124	0.23
g_Lactobacillus	26	0.00	0.00	6720.23	0.45
o_Caulobacterales	26	0.00	0.00	Inf	0.58
c_Flavobacteriia	26	2.60E+21	0.00	1.31E+46	0.09	25	3.81E+60	1.62E+19	8.99E+101	0.004
g_Corynebacterium	26	0.15	0.00	459025.31	0.80
g_Halomonas	26	0.00	0.00	3.33E+07	0.32
f_Micrococcaceae	26	9.36	0.13	654.68	0.30
g_Parvimonas	26	0.00	0.00	4.36E+67	0.51

FIG 4.

Predictors of overall survival at the end of treatment (EOT). FIGO, International Federation of Gynecology and Obstetrics.

DISCUSSION

In this study, we characterized the cervical microbiome in WLWH and HIV-negative women with cervical cancer from Botswana, a country with a high HIV prevalence and a significant cervical cancer burden. Our findings reveal notable differences in microbiota composition related to HIV status and demonstrate the potential prognostic value of specific microbial taxa in predicting survival outcomes in WLWH treated with CRT for cervical cancer. Our findings provide further evidence that the cervical tumor microenvironment, particularly in women with HIV infection, plays a crucial role in shaping microbiota dynamics and influencing survival outcomes. These results suggest the need for a better understanding of regional and population-specific microbial dynamics in cancer care, particularly in areas with unique disease burdens like Botswana.

HIV Status and Cervical Microbiome Composition

We identified distinct microbial patterns in WLWH compared with HIV-negative women. The WLWH exhibited enrichment of the families Micrococcaceae and Corynebacteriaceae and the genus Corynebacterium, while the HIV-negative women had higher levels of the family Gemellaceae and genus Nesterenkonia. These findings highlight how the immunosuppressed state in WLWH influences the cervical microbiome.^19–21 In Botswana, where HIV prevalence is among the highest globally,²² our findings provide critical insights into the interplay between HIV infection, immune status, and cervical cancer pathogenesis. The lower Pielou evenness in WLWH suggests a more imbalanced microbiota, which could contribute to both the increased risk and aggressive nature of cervical cancer observed in WLWH in high-HIV-prevalence regions and contribute to the pathogenesis and progression of cervical cancer in this population.^{23, 24} This imbalance could also reflect the impact of HIV on local immune defenses, which are essential for maintaining microbial homeostasis in the cervical environment.^25–27 Additionally, not surprisingly in this cervical cancer cohort, regardless of HIV status, patients exhibited a high prevalence of CST IV-C. CST IV-C is classically marked by diverse, non-Lactobacillus-dominated microbiota including Gardnerella vaginalis and Atopobium vaginae, indicative of a dysbiotic state which has been linked to bacterial vaginosis and the development of cervical dysplasia and cancer. A higher percentage of HIV-negative women than WLWH exhibited CST I, characterized by dominance of Lactobacillus crispatus, which is typically associated with a healthy, low-diversity vaginal microbiome.

Changes in Cervical Microbiome Between Baseline and End of CRT in WLWH

We observed that although overall cervical tumor microbiome diversity remained stable in WLWH undergoing CRT, the composition of the microbiome shifted. Before CRT, the families Alcaligenaceae and Lactobacillaceae were enriched, while after treatment, Ruminococcus and Caulobacterales were more abundant. Given that in sub-Saharan African countries like Botswana, cervical cancer is highly prevalent and often is diagnosed in women who are also immunocompromised due to HIV, the resilience of the microbiome during CRT may reflect the body’s attempt to maintain homeostasis in a highly stressed environment.²⁸ Lactobacillus species, which are normally dominant in a healthy vaginal and cervical microbiome, play a key role in maintaining microbial homeostasis by producing lactic acid, which lowers pH and inhibits pathogenic bacteria.²⁹ In our study, enrichment of Lactobacillaceae was noted in WLWH before CRT. However, there was a decrease in the proportion of women with Lactobacillus-dominant microbiome after treatment potentially indicating that CRT disrupts the protective role of Lactobacillus in the cervical environment. Additionally, studies indicate that tumor-resident Lactobacillus iners confers resistance to CRT through lactate-induced metabolic rewiring.³⁰ In WLWH, in whom immune dysregulation is common, the initial presence of Lactobacillus may act as a residual protective factor, but its decline after treatment may contribute to microbial imbalances. These imbalances could foster a more pro-inflammatory environment, negatively impacting survival outcomes. The shift toward Ruminococcus and Caulobacterales after treatment might be linked to metabolic changes induced by CRT or alterations in immune responses specific to WLWH undergoing cancer treatment in Botswana and may be associated with treatment response.

Microbiota as Predictors of OS in WLWH Undergoing CRT

Flavobacteriia was identified as an independent predictor of OS in WLWH undergoing concurrent CRT. Flavobacteriia, previously referred to as Flavobacteria,²⁹ is the largest class within the phylum Bacteroidetes, comprising families such as Flavobacteriaceae, Blattabacteriaceae, and Cryomorphaceae. Its presence in the vaginal microbiota is indicative of dysbiosis and reflects an atypical or altered microbiome composition, as Flavobacteriia are not typically dominant in healthy, Lactobacillus-dominated vaginal environments, such as those characterized by CST I and CST III. Although Flavobacteriia are primarily recognized for their ecological importance, recent evidence suggests a potential association with human cancers. In studies investigating the oral microbiome, Flavobacteriia were enriched in extranodal extension–negative oral squamous cell carcinomas.³¹ Similarly, in lower-lobe non–small cell lung cancer, peritumoral tissues were enriched in Flavobacteriia, with an inverse abundance relationship to Actinobacteria in extratumoral tissues, varying by tumor lobe location.³² These findings underscore the need for further investigation of the role of Flavobacteriia in cancer pathophysiology. Our results suggest that Flavobacteriia may represent a novel microbial target for therapeutic intervention, particularly for improving survival outcomes for WLWH in Botswana, a setting where HIV and cervical cancer pose significant public health challenges.³³

Clinical Implications for Botswana and Sub-Saharan Africa

Our findings have several possible important clinical implications for cervical cancer management in Botswana and other sub-Saharan African countries with high HIV burdens. First, the stability of microbial diversity during CRT suggests that interventions targeting specific harmful taxa, rather than global microbial shifts, could enhance treatment outcomes. Second, identifying microbial markers such as Flavobacteriia offers a noninvasive means of predicting survival and tailoring treatment strategies in this population. It is possible that taxa like Flavobacteriia inhabiting the cervical microbiome may be manipulated to improve treatment response. Knowing specific cervical microbial organisms that inhabit and change during CRT provides further insight into mechanisms that may modulate immune response and potentiate treatment outcomes.

Given the resource constraints in Botswana and similar regions, integrating microbiome-based diagnostics could aid in the development of more personalized and effective cancer care strategies.^34–36 Additionally, understanding how HIV infection alters the cervical microbiome in women with cervical cancer could lead to new interventions aimed at reducing the overall burden of cervical cancer in WLWH, a substantial proportion of the women in southern Africa.

Limitations and Future Directions

Although our study provides important insights, it was completed in a resource limited setting and the relatively small sample size and focus on a specific geographic region may limit the generalizability of our findings. The HIV-negative cohort is small limiting the statistical power of between-group comparisons. Furthermore, information was not available on anti-retroviral use, CD-4 levels, treatment delays, and ECOG performance status in our data set which may have an impact on the findings. Additionally, specific radiologic response measurements for each patient were not available at the time of data analysis. However, given Botswana’s high HIV prevalence, our results are highly relevant to understanding cervical cancer progression in WLWH in Botswana and similar populations. Further research is needed to explore microbiome dynamics across different stages of cancer treatment and in populations with different HIV prevalence. Additionally, future studies should investigate the functional roles of the identified microbial taxa in modulating immune responses, especially in people living with HIV. Such insights could inform the development of novel therapies tailored to the unique microbial and immunological environment in cervical cancer patients in high-HIV-burden regions like Botswana.

Conclusion

In conclusion, our study provides important insights into the cervical microbiome of WLWH compared with HIV-negative women with cervical cancer in Botswana, revealing distinct microbial profiles influenced by HIV status and revealing the dynamics of the microbiome during CRT. The identification of Flavobacteriia as an independent predictor of OS in WLWH highlights the potential role of the microbiome in influencing cancer outcomes, particularly in regions with high HIV prevalence. Furthermore, the baseline enrichment and subsequent decline of Lactobacillus in WLWH suggests that microbiome disruptions during treatment may impact prognosis. Our study demonstrates hypothesis-generating differences in cervical microbial profiles of WLWH compared with HIV-negative women with cervical cancer in Botswana. . These findings provide rationale for further study of the cervical microbiome in cervical cancer. Future research should focus on further elucidating the functional roles of key microbial species in WLWH with cervical cancer.

Context Summary:

Key Objective

Can distinct features of the cervical microbiome in women living with HIV (WLWH) in Botswana predict treatment and survival outcomes in cervical cancer?

Knowledge Generated

Women living with HIV in Botswana exhibited a more imbalanced and potentially pathogenic cervical microbiome compared to HIV-negative women. In WLWH, specific microbial shifts after chemoradiation—particularly increased Flavobacteriia—were associated with worse overall survival.

Relevance

These findings suggest that cervical microbiome composition may have prognostic value in WLWH in Botswana with cervical cancer. Further investigation could clarify whether microbiome profiling has a role in guiding treatment strategies