Prevalence of High-Risk Nonvaccine Human Papillomavirus Types in Advanced Squamous Cell Carcinoma Among Individuals of African vs Non-African Ancestry.

Introduction

High-risk human papillomavirus (hrHPV) causes squamous cell carcinoma (SCC) at a variety of sites, including cervix, head and neck, anus, vulva, vagina, and penis. The 9-valent HPV vaccine is highly effective at preventing carcinomas at these sites related to the specific HPV types it covers. Among women with hrHPV detected in routine cervical swabs, however, the distribution of hrHPV types is known to differ by racial/ethnic group, with East and West Africans[1] and North Americans of African ancestry[2,3] showing overrepresentation of hrHPV types that are not covered by the current 9-valent HPV vaccine. Corresponding studies of HPV type distribution by ethnicity in advanced-stage hrHPV-related cancers are limited. In this investigation, we performed a retrospective analysis of a large multiethnic study of advanced-stage carcinoma samples to test the hypothesis that patients of African vs non-African ancestry may show dissimilar distributions of nonvaccine hrHPV types.

Methods

Our archive of 290 311 tumor samples, each from a different patient, underwent comprehensive genomic profiling using a hybrid-capture–based DNA sequencing platform.[4] The samples were sent from medical care facilities across North America from January 2014 to June 2020 for detection of targetable genetic alterations during routine clinical care. Western Institutional Review Board approved this study, including issuing an informed consent waiver and a HIPAA waiver of authorization. This report follows the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cross-sectional studies. We detected hrHPV genomic sequences in 3793 cervical, anogenital, and head and neck SCCs by de novo assembly of nonhuman reads, followed by alignment to the RefSeq database, which distinguishes hrHPV types.[5] Patient ancestry was determined by classifying specific single-nucleotide variations by genomic profiling based on their known variation among populations in the 1000 Genomes Project.[6] The comparator groups were individuals of African vs non-African ancestry. Categorical data were analyzed by Fisher exact test using Prism version 8.3.1 (GraphPad Software). Data were analyzed in July 2020. A 2-tailed P value of <.05 was considered statistically significant; the Bonferroni correction was applied for multiple simultaneous comparisons. Statistical analysis was performed for hrHPV types with 10 or more cancer cases total in individuals of African ancestry.

Results

We identified 5 hrHPV types each with 10 or more SCC cases in individuals of African ancestry: HPV-16 (193 cases), HPV-18 (56 cases), HPV-35 (17 cases), HPV-45 (27 cases), and HPV-59 (16 cases). Compared with non-African groups, individuals of African ancestry showed significant, several-fold enrichments for HPV types 35 (5.3% vs 1.6%; P < .001), 45 (8.4% vs 3.4%; P < .001), and 59 (5.0% vs 0.5%; P < .001) (Figure), with significant enrichments in cervical (HPV-35, 5.2% vs 1.3%; P = .003; HPV-59, 9.0% vs 0.9%; P < .001) and anal (HPV-45, 7.0% vs 1.0%; P = .003) carcinomas after stratification by primary site (Figure).

Figure.

Comparative Frequencies of High-Risk Human Papillomavirus (hrHPV) Types Among hrHPV-Related Squamous Cell Carcinomas, Stratified by African vs Non-African Patient Ancestry

Displayed are data for hrHPV types affecting more than 10 patients of African ancestry in our multiethnic tumor archive. The Bonferroni correction for 5 multiple simultaneous comparisons was applied for each analysis; significant P values (<.05/5 = .01) are indicated within each panel. In panel A, a total of 3793 pan-hrHPV–related squamous cell carcinomas (321 [8%] from patients of African ancestry) were included for analysis. Stratified by primary site, panel B (cervix) included 1255 carcinomas (156 [12%] from patients of African ancestry); panel C (anus) included 865 carcinomas (71 [8%] from patients of African ancestry); panel D (head and neck) included 1363 carcinomas (49 [4%] from patients of African ancestry); panel E (vulva) included 147 carcinomas (19 [13%] from patients of African ancestry); panel F (vagina) included 91 carcinomas (15 [16%] from patients of African ancestry); and panel G (penis) included 72 carcinomas (11 [15%] from patients of African ancestry).

The 2 nonvaccine hrHPV types (HPV-35 and HPV-59) accounted for 10.3% (33/321 cases) of all hrHPV-related SCCs in persons of African ancestry, but only 2.1% (72/3472 cases) of hrHPV-related SCCs in individuals of other ancestries, a difference of nearly 5-fold (P < .00001). Prevalences were comparatively low across all 4 non-African ancestral groups (eTable in the Supplement). Demographic data by primary site and African vs non-African ancestry are provided in the Table. Tumor mutational burden and genomic alterations showed essentially no significant differences based on patient ancestry, suggesting similar tumorigenic mechanisms (Table).

Table.

Comparisons of Demographic Characteristics, Tumor Mutational Burden, Microsatellite Instability, and Gene-Specific Pathogenic Alterations in hrHPV-Related Squamous Cell Carcinomas, Stratified by Primary Site and African vs Non-African Patient Ancestry

Characteristics	Cervix			Anus			Head and neck			Vulva			Vagina			Penis
	African	Non-African	P value	African	Non-African	P value	African	Non-African	P value	African	Non-African	P value	African	Non-African	P value	African	Non-African	P value
No. of cases	156	1099	NA	71	794	NA	49	1314	NA	19	128	NA	15	76	NA	11	61	NA
Female, %	100	100	NA	75	75	>.99	12	11	>.99	100	100	NA	100	100	NA	0	0	NA
Male, %	0	0	NA	25	25	>.99	88	89	>.99	0	0	NA	0	0	NA	100	100	NA
Age, median (range), y	49.5 (16-84)	47.0 (22-84)	.19	57 (31-91)	60 (29-96)	.02	58 (20-81)	60 (8-96)	.046	53 (34-83)	60 (21-91)	.07	56 (35-94)	61 (25-96)	.13	57 (43-73)	67 (24-86)	.009
TMB, median (IQR), mutations/Mb	5.0 (3.5-8.7)	5.0 (2.6-9.2)	.49	6.3 (3.5-8.8)	4.4 (2.6-7.5)	.007	5.0 (3.5-10.0)	3.5 (2.1-7.0)	.004	5.0 (2.5-7.5)	5.0 (2.6-8.8)	.70	5.0 (3.2-10.8)	5.0 (3.5-11.7)	.71	5.2 (3.8-11.3)	5.2 (2.6-11.3)	.40
MSI high, %	1	1	.71	1	0.7	.45	0	1	>.99	0	2	>.99	0	2	>.99	0	2	>.99
Gene, % alteration
PI3K/AKT/mTOR pathway
PIK3CA	30	36	.15	37	33	.51	22	27	.52	21	31	.43	27	40	.40	64	33	.09
KMT2D	10	17	.05	20	19	.75	12	12	>.99	16	16	>.99	13	20	.73	9	23	.44
PTEN	10	11	.59	9	16	.09	31	13	.002	5	14	.47	7	13	.68	0	10	.58
FBXW7	7	12	.06	16	12	.45	18	9	.04	11	9	>.99	0	12	.35	9	8	>.99
STK11	11	10	.78	9	4	.14	2	4	>.99	5	10	.70	7	9	>.99	0	8	>.99
SOX2	4	6	.45	7	10	.67	6	5	.75	0	4	>.99	0	9	.60	0	0	>.99
Epigenetic regulation
KMT2C	5	8	.33	24	14	.02	11	10	.81	0	8	.36	0	12	.35	13	16	>.99
EP300	8	9	>.99	9	7	.46	12	5	.05	5	13	.47	0	15	.20	0	13	.34
BAP1	3	6	.07	4	2	.42	0	1	>.99	0	6	.60	7	8	>.99	0	0	>.99
DNA damage
TP53	7	6	.59	4	3	.73	10	9	.62	5	6	>.99	0	8	.58	9	10	>.99
Single pass transmembrane receptor
NOTCH1	7	4	.04	1	4	.35	16	7	.02	5	6	>.99	0	4	>.99	0	8	>.99
Cell cycle regulation
CCND1	1	2	>.99	3	4	>.99	6	3	.19	0	3	>.99	7	3	.42	18	5	.17
CDKN2A	3	1	.51	1	0.8	.45	12	4	.02	0	2	>.99	0	0	>.99	0	10	.58
TERT promoter	11	13	.61	0	1	>.99	2	2	>.99	6	7	>.99	0	9	.60	13	12	>.99
RB1	5	6	.72	7	4	.20	2	7	.25	11	4	.22	7	4	.52	9	2	.28
Receptor tyrosine kinase
FGFR3	3	4	.51	6	4	.52	2	5	.51	5	2	.43	7	7	>.99	0	3	>.99
ERBB2	7	5	.36	3	2	.38	2	2	>.99	0	0.8	>.99	0	3	>.99	9	0	.15
EGFR	4	3	.62	3	2	.38	0	2	>.99	0	2	>.99	0	0	>.99	0	5	>.99
RAS/MAPK pathway
KRAS	2	4	.47	0	3	.40	2	2	.57	0	3	>.99	0	7	.59	0	5	>.99
HRAS	0	0.7	.61	1	0.4	.29	2	2	.56	0	2	>.99	7	3	.42	0	2	>.99
BRAF	1	1	>.99	0	1	>.99	0	0.8	>.99	0	0.8	>.99	0	1	>.99	0	2	>.99
NFκB pathway
CYLD	1	0.9	>.99	5	13	.09	4	17	.02	0	5	>.99	0	3	>.99	0	0	>.99
Apoptosis
CASP8	7	4	.04	9	2	.01	8	2	.02	0	6	.60	0	4	>.99	9	8	>.99
Immune regulation
PDL1	3	3	>.99	2	3	>.99	2	3	>.99	6	0.8	>.99	7	0	.17	0	7	>.99
PDL2	3	3	.80	2	2	>.99	2	2	.57	0	0.8	>.99	0	0	>.99	0	8	>.99
Transcription factor (TGF-B signaling)
SMAD4	7	2	<.001b	0	0.8	>.99	2	1	.40	0	2	>.99	7	3	.42	0	2	>.99
Transcriptional regulation
AR	0	0.3	>.99	0	0.3	>.99	0	0.1	>.99	5	2	>.99	0	0	>.99	18	2	.06

Abbreviations: hrHPV, high-risk human papillomavirus; IQR, interquartile range; Mb, megabase; MSI, microsatellite instability; mTOR, mammalian target of rapamycin; NA, not applicable; TMB, tumor mutational burden.

The Mann-Whitney U test was used for comparisons of age and TMB. The Bonferroni correction for 31 multiple simultaneous comparisons was applied.

Denotes the sole significant P value (<.05/31 = .0016).

Discussion

In our pan-hrHPV analysis of SCCs, hrHPV types not covered by the current 9-valent vaccine were increased in individuals of African ancestry. It has been hypothesized that HPV-35 might be better adapted to individuals of African ancestry,[3] or the increased prevalence may be a consequence of ancient geographic isolation.[3] The current study was limited by its retrospective nature, lack of knowledge of vaccination status, and the nature of our underlying sample set, which favored advanced disease. While our study indicated that the large majority of hrHPV types identified in advanced SCCs are included in the current 9-valent vaccine irrespective of patient ancestry, our data argue for the addition of HPV types 35 and 59. Future studies and vaccine decisions should involve consideration of ancestrally diverse populations.