Abstract
Background:
People with breast cancer experience high symptom burden from the psychoneurological symptom cluster (fatigue, anxiety, depression, sleep disturbances, and pain). However, interventions to manage this symptom cluster are limited as its etiology is not well understood. One system that has been linked with this symptom cluster in both healthy and other chronically ill populations but has not been explored in breast cancer is the oral microbiome.
Objective:
This study examined differences in oral microbiome diversity and composition between individuals with breast cancer who reported high versus low levels of psychoneurological symptoms.
Methods:
Oral microbial changes based on bacterial diversity and relative abundance were estimated from 16S rRNA gene sequencing data of oral tongue swabs from 22 individuals with breast cancer prior to mastectomy. The fatigue, anxiety, depression, sleep disturbance, and pain interference subscales of the PROMIS-29 scale were used to calculate the symptom groups.
Results:
Observed amplicon sequence variants were statistically lower in the high symptom group than in the low symptom groups but no differences between groups were seen with the other alpha diversity metrics (Shannon, Inverse Simpson). No differences in beta diversity were found between the high and low symptom groups. For differential abundance, three genera were more abundant in the high symptom group than the low symptom group, Aggregatibacter, Lautropia, and Prevotellaceae UCG-001.
Discussion:
Preliminary associations were identified between the psychoneurological symptom cluster and oral microbiome diversity and composition in our cohort of people with breast cancer. Of the differentially abundant genera, Lautropia has been previously linked with anxiety and Aggregatibacter has been researched studied due to its pro-inflammatory potential. As inflammation is posited to influence psychoneurological symptom cluster severity, further research on the relationship between the oral microbiome and the psychoneurological symptom cluster, specifically in relation to potential causal mechanisms, is warranted. Enhanced insight into this relationship could inform targeted interventions to improve the symptom experience for people with breast cancer.
Keywords: oral microbiome, breast cancer, psychoneurological symptoms
INTRODUCTION
Breast cancer negatively affects quality of life and is associated with a wide array of debilitating symptoms (Khajoei et al., 2023). Frequently reported symptoms for people with breast cancer include fatigue, depression, stress, and pain (Khajoei et al., 2023). However, most symptoms do not occur in isolation; it is common for more than one symptom to be experienced by people with breast cancer (Li, 2021). This co-occurrence of two or more interrelated symptoms is referred to as a symptoms cluster (Li, 2021). One group of frequently co-occurring symptoms that are particularly distressing for people with breast cancer is the psychoneurological symptom cluster, which is composed of fatigue, depression, anxiety, pain, and sleep disturbances (Grayson et al., 2023; Tometich et al., 2019). Various studies have demonstrated that these symptoms are both prevalent as well as cluster together in people with breast cancer prior to treatment (Doong et al., 2015; Tometich et al., 2019), such as surgery or chemotherapy, as well as across the treatment trajectory (Grayson et al., 2023). A notable large cohort study in older individuals with breast cancer found that 16% of participants had high burden from the psychoneurological symptom cluster prior to any treatment (Tometich et al., 2019). Yet, while research has shown that the psychoneurological symptom cluster is highly prevalent and negatively affects quality of life for people with breast cancer, psychoneurological symptom cluster management is challenging (Miaskowski et al., 2017).
Effectively managing symptoms is critical to ensuring that people with breast cancer maintain a high quality of life. Conversely, poor symptom management can lead to cessation of treatment, leading to worse outcomes and prognosis (Kidwell et al., 2014; Richardson et al., 2007). While the importance of symptom management is well-established, interventions to manage symptoms, including those within the psychoneurological symptom cluster, do not sufficiently reduce symptom burden for all people with breast cancer (Miaskowski et al., 2017; Starkweather et al., 2013). One reason why symptom management interventions are not always effective is because the underlying mechanisms contributing to psychoneurological symptom development and severity in people with breast cancer are not clearly understood (Starkweather et al., 2013). Multiple pathways have been shown to contribute to the development of psychoneurological symptoms, such as dysregulation of the hypothalamic-pituitary adrenocortical axis, inflammation, and oxidative stress, (Lyon et al., 2014; Starkweather et al., 2013). However, personalized interventions are still inadequate as the previously explored pathways do not fully explain the variation in psychoneurological symptom cluster development and severity experienced by people with breast cancer. Thus, more research is needed to better understand the underlying mechanisms which lead to high psychoneurological symptom burden in people with breast cancer (Lyon et al., 2014; Miaskowski et al., 2017; Starkweather et al., 2013).
One biological system that has been preliminarily linked with various symptoms within the psychoneurological symptom cluster is the oral microbiome (Qiu et al., 2025; Simpson et al., 2020; Wingfield et al., 2021). The oral microbiome is the second largest community of microbes within the human body, behind the gut, and has gained increasing attention due to its connection with numerous chronic conditions, including cancer, and overall well-being (Rajasekaran et al., 2024). Additionally, the oral microbiome has been linked to multiple psychoneurological symptoms, such as fatigue and depression, in both healthy and chronically ill populations (Lupo et al., 2021; Peng et al., 2025; Qiu et al., 2025; Wingfield et al., 2021). Furthermore, research has shown that specific oral microbes can promote inflammation (Moreno et al., 2023; Rajasekaran et al., 2024), a previously identified factor associated with increased psychoneurological symptom cluster burden (George et al., 2020; Starkweather et al., 2013).
Due to the high psychoneurological symptom burden in people with breast cancer and the link between psychoneurological symptoms and the oral microbiome in other populations, the purpose of this study was to examine differences in oral microbiome diversity and composition between individuals with breast cancer who report high versus low levels of psychoneurological symptoms. The microbiome is commonly evaluated by examining community characteristics, such as the richness and evenness of a community (alpha diversity) as well as the similarity/dissimilarity of communities between samples (beta diversity) (Kers & Saccenti, 2022). The microbiome is also frequently assessed at a microorganism level through examining the proportional representation of taxon abundance within a sample comparative to the entire microbial community of the sample (relative abundance). As the oral microbiome is a biological system that may contribute to psychoneurological symptom burden for people with breast cancer, better insight into this relationship between the oral microbiome and the psychoneurological symptom cluster could help inform future symptom management interventions. Specifically, if the oral microbiome is identified to causally affect psychoneurological symptom cluster severity, targeted interventions to change oral microbiome composition, such as dietary or dental hygiene interventions, could be utilized to reduce symptom cluster severity in people with breast cancer.
METHODS
Design And Sample
This cross-sectional study examined the association between the oral microbiome and the psychoneurological symptom cluster in 22 individuals with breast cancer prior to a mastectomy and chemotherapy. Demographic and clinical data as well as biospecimens to characterize oral microbiome composition for this study was obtained through the 1,000 Patient Project (1KP) biorepository. 1KP is a comprehensive biospecimen and data collection repository of individuals who have received or are scheduled to receive surgery at a comprehensive cancer center in the southeastern United States. Individuals in the biorepository provide informed consent for comprehensive data and biospecimen collection at multiple timepoints pre- and post-operatively over the course of 6 months. The biorepository includes individuals who meet the following eligibility criteria: (1) > 18 years old, (2) undergoing a surgical procedure, and (3) agree to provide a blood and/or urine sample both pre- and post-procedure. Other collections, such as oral tongue swab samples, were optional.
In addition to the biorepository eligibility described above, additional inclusion criteria for this study were: (1) diagnosed with in situ or invasive breast cancer (identified through self report then verified with ICD-10 codes), (2) scheduled for a mastectomy, (3) no documentation of neoadjuvant chemotherapy at preoperative visit, (4) participated in oral tongue swab collection at preoperative visit, and (5) completed the Patient-Reported Outcomes Measurement Information System® Profile 29 (PROMIS-29) questionnaire at preoperative visit. Of note, the 1KP biorepository includes both a preoperative initial encounter visit and a surgery preoperative visit; individuals who completed both the oral swab collection and PROMIS-29 questionnaire at either preoperative timepoint were included in this study. This nested study within the 1KP biorepository was approved by the Duke University Health System Institutional Review Board.
Data Collection
Study participants received a Research Electronic Data Capture (REDCap) questionnaire, part of the REDCap electronic data capture tools hosted at the academic center, sent via email to provide data on demographics, health history, and their symptom experience (Harris et al., 2009). Participants were given the 1KP study team contact information and instructed to contact the study team if they had any questions or concerns. Additional data, such as cancer stage, cancer diagnosis, major psychiatric illness history, and history of radiation, were manually extracted by research coordinators via electronic health record review. For major psychiatric illness history, participants were recorded as having a condition if they had a history of any disorder found in the Diagnostic and Statistical Manual of Mental Disorders, 5th Edition, excluding alcohol use disorder and substance (drug) use disorder.
Symptom Measurement
The PROMIS-29 questionnaire was used to determine psychoneurological symptom scores. The PROMIS-29 consists of 29-items which measure symptom severity on a 5-point Likert scale. The five symptoms examined in this study were fatigue, anxiety, depression, sleep disturbances, and pain interference, all which have 4-item subscales. For these subscales higher scores indicated higher symptom severity. To aide in interpretation, PROMIS-29 subscales are calculated into T-scores with a population average of 50 (SD=10). Symptom scores between 55–60 are mild, between 60–70 are moderate, and between 70–80 are severe. To create symptoms groups, individuals with two or more symptoms (out of the five examined) at or above the mild threshold were grouped as high symptom burden and those with fewer than two mild symptoms were classified as low symptom burden. The cut-off of two symptoms at a mild or worse severity was chosen to align with the definition of a symptom cluster that states a cluster requires two or more co-occurring symptoms. Reliability, validity, and test –retest scores are high and acceptable (Cella et al., 2019; Hays et al., 2018).
Oral Sample Collection
The 1KP biorepository collected oral tongue swab samples from participants using the OMNIGene ORAL OMR-120 collection device (DNAGenotek, Ontario, Canada). Participants were given printed instructions for the tongue collection at home which instructed participants to rub the tongue with the swab for a minimum of 30 seconds, return the swab to the device, and close tightly. After collection, participants returned the room temperature swabs in-person or via FedEx shipment. Upon receipt, samples were processed based on the manufacturer’s recommendations for preparation and storage. In brief, samples were vortexed, and the swabs were pressed against the collection tube’s inner wall to remove excess liquid. Proteinase K was then added to the sample collection tubes and the tubes were vortexed again and incubated at 50°C for one hour in a water bath. Post incubation, the samples were aliquoted into cryovials and frozen at −80°C.
Oral Sample Processing
Oral sample aliquots from qualifying individuals were identified from the 1KP biorepository and transferred to the Duke Microbiome Core. DNA extraction was completed using the Qiagen DNeasy PowerSoil Pro HT DNA extraction kit on the QIACube using standard protocols. In short, frozen samples were thawed and 250ul of the sample was added to PowerSoil bead tube with CD1 lysis buffer. The tube was then vortexed and transferred to a heat block at 65°C for 10 minutes. The tubes were then loaded onto vortex genie and agitated for 5 minutes to facilitate mechanical lysis of bacteria within solution. Lysate was centrifuged and 500–600μl of supernatant was transferred to a new tube with CD2 solution. After mixing, the samples were loaded and run on the QIAcube HT instrument using the DNeasy 96 Powersoil Pro QIAcube HT. After loading, the Qiacube transferred samples to CD3 buffer, mixed, loaded mixture onto a silica column to bind DNA. Using a vacuum, the loaded solution was pulled through the silica column. The column was washed with AW1 wash buffer followed by a second wash with AW2 buffer to remove residual salts from the lysis and binding buffers. The column was then dried, and elution buffer was loaded and run through the column into collection tubes.
Post extraction, DNA was quantified via Qubit and assessed for purity via NanoDrop prior to normalization to 1.0ng/ul and preparation of 16S rRNA gene amplicon sequencing libraries. The 16S rRNA gene variable region 4 primer set (515F-806R) of the normalized data were then amplified as outlined in the Earth Microbiome Project 16S rRNA amplicon sequencing protocol (Apprill et al., 2015; Caporaso et al., 2012). The barcoded libraries were cleaned using AMpure XP PCR purification beads, quantified via Qubit, and pooled in equimolar concentration for sequencing. Pooled libraries were loaded onto Illumina MiSeq instrument and run through a paired end 250 cycle program to generate 250 base-pair reads.
Bioinformatic Processing
Sequencing reads were demultiplexed, quality filtered, denoised, and trimmed using DADA2 pipeline (Callahan et al., 2016). After chimera removal, the SILVA v138 99% database was used for taxonomic classification. Non-bacteria sequences identified through taxonomic classification were removed. The data was then run through the DECONTAM pipeline and bacteria above a 0.15 threshold of likelihood to be a contaminant were removed (Davis et al., 2018). After cleaning, the average read count was 46,068 with a range of 6,912 to 75,296; no samples were filtered out due to low read count (<5,000 reads). Cleaned data was analyzed using the R studio “phyloseq” package to estimate alpha diversity (observed amplicon sequence variants (ASVs), Shannon diversity, Inverse Simpson diversity) and beta diversity (Jaccard and Bray-Curtis) in addition to the genus-level relative abundances of genera with greater than 10% prevalence (McMurdie & Holmes, 2013). We included multiple alpha diversity metrics as they estimate unique characteristics of richness and evenness; similarly, both Jaccard and Bray-Curtis were calculated as Jaccard strictly uses presence/absence data while Bray-Curtis incorporates bacterial relative abundance.
Quality Control
Two negative extraction controls were included in this study. Importantly, both negative controls were removed during the DADA2 pipeline. A positive control was also utilized to ensure successful DNA extraction, but the positive control was not sequenced.
Statistical Analysis
Descriptive statistics were used to summarize overall participant characteristics and by high or low psychoneurological symptom cluster. Differences in participant characteristics and symptom severity between the symptom cluster groups were tested using Welch’s t-test or Wilcoxon’s rank-sum test for continuous variables and Fisher’s exact test for categorical variables.
Boxplots were created to visualize the alpha diversity distributions for the high and low symptom groups. A linear model including adjustment for sequencing depth was used to estimate differences in observed ASVs, Shannon diversity, and Inverse Simpson diversity between the high and low symptom groups. Differences in the high and low symptom groups in the beta diversity matrices (Bray-Curtis and Jaccard) were tested using permutational analysis of variance with adjustment for sequencing depth (PERMANOVA). Principal coordinates analysis (PCoA) vectors were generated and presented to visualize differences in oral microbial communities between the symptom groups. The between-group demographic and symptom comparisons, linear models, and PERMANOVA testing used a set significance threshold of p ≤ .05.
The differential abundance method, LinDA (linear regression framework for differential abundance analysis, was used to robustly identify genera (with a prevalence of greater than 10%) associated with symptom groups (Zhou et al., 2022); p-values were adjusted for multiple comparisons to control the false discovery rate. A p-adjusted value of p-adj ≤ .05 was considered statistically significant.
Smoking status (never smoked versus former smoker), alcohol use (never/former drinker versus current drinker), and age were controlled for in this study due to their ability to modify the oral microbiome (Carbone et al., 2025; Chattopadhyay et al., 2024; Fan et al., 2018). R version 4.5.1 statistical software program was used for data analysis.
RESULTS
The demographic and clinical data of the participants overall and by symptom group is shown in Table 1. Twenty-two individuals living with breast cancer were included in this study. Of note, 23 individuals met the inclusion and exclusion criteria but as only one participant reported that they were a current smoker, they were removed from all analysis to ensure the analyses were not adversely affected by the small singular sample. While gender was not an inclusion criterion, all participants in this study were female. The average age was 57.5 (SD= 13.0) years, and most participants reported identifying as non-Hispanic/Latino and White. Of importance, the high symptom group had a large proportion of participants with a history of or current diagnosis of a major psychiatric condition, with 10 of the 11 people (90.91%) with this condition categorized to the high symptom group.
Table 1.
Demographic and Clinical Data
| Low Symptoms (N=10) | High Symptoms (N=12) | Overall (N=22) | p-value | |
|---|---|---|---|---|
| Age | ||||
| Mean (SD) | 61.0 (14.1) | 54.7 (11.8) | 57.5 (13.0) | .2739 |
| Median [Min, Max] | 61.0 [38.0, 80.0] | 56.5 [40.0, 73.0] | 56.5 [38.0, 80.0] | |
| Sex | ||||
| Female | 10 (100.0%) | 12 (100.0%) | 22 (100.0%) | 1.0000 |
| BMI | ||||
| Mean (SD) | 25.9 (3.40) | 31.6 (10.5) | 29.4 (8.81) | .3283 |
| Median [Min, Max] | 25.0 [21.4, 30.4] | 29.5 [19.4, 52.3] | 27.3 [19.4, 52.3] | |
| Missing | 3 (30.0%) | 1 (8.3%) | 4 (18.2%) | |
| Race | ||||
| White | 7 (70.0%) | 9 (75.0%) | 16 (72.7%) | .8010 |
| Black or African American | 2 (20.0%) | 3 (25.0%) | 5 (22.7%) | |
| Unknown/Not Reported | 1 (10.0%) | 0 (0%) | 1 (4.5%) | |
| Ethnicity | ||||
| Hispanic/Latino | 1 (10.0%) | 0 (0%) | 1 (4.5%) | .4545 |
| Non-Hispanic/Non-Latino | 9 (90.0%) | 12 (100%) | 21 (95.5%) | |
| Primary Insurance Type | ||||
| Private/Commercial | 4 (40.0%) | 7 (58.3%) | 11 (50.0%) | .5027 |
| Medicare | 5 (50.0%) | 3 (25.0%) | 8 (36.4%) | |
| Other Government | 1 (10.0%) | 2 (16.7%) | 3 (13.6%) | |
| Smoking Status | ||||
| Never Smoked | 7 (70.0%) | 6 (50.0%) | 13 (59.1%) | .4149 |
| Former Smoker | 3 (30.0%) | 6 (50.0%) | 9 (40.9%) | |
| Alcohol Consumption * | ||||
| Current Drinker | 5 (50.0%) | 9 (75.0%) | 14 (63.6%) | .3777 |
| No at Time of Encounter | 3 (30.0%) | 3 (25.0%) | 6 (27.3%) | |
| Never Drank | 2 (20.0%) | 0 (0%) | 2 (9.1%) | |
| Radiation Therapy | ||||
| No | 4 (40.0%) | 8 (66.7%) | 12 (54.5%) | .6267 |
| Yes | 3 (30.0%) | 3 (25.0%) | 6 (27.3%) | |
| Missing | 3 (30.0%) | 1 (8.3%) | 4 (18.2%) | |
| History of/Current Major Psychiatric Condition | ||||
| No | 9 (90.0%) | 2 (16.7%) | 11 (50.0%) | .0019** |
| Yes | 1 (10.0%) | 10 (83.3%) | 11 (50.0%) | |
| Cancer Stage | ||||
| 0 | 2 (20.0%) | 2 (16.7%) | 4 (18.2%) | .5769 |
| Ia | 3 (30.0%) | 5 (41.7%) | 8 (36.4%) | |
| Ib | 0 (0%) | 2 (16.7%) | 2 (9.1%) | |
| IIa | 1 (10.0%) | 1 (8.3%) | 2 (9.1%) | |
| IIb | 2 (20.0%) | 0 (0%) | 2 (9.1%) | |
| IIIb | 0 (0%) | 1 (8.3%) | 1 (4.5%) | |
| Missing | 2 (20.0%) | 1 (8.3%) | 3 (13.6%) |
Former/never drinkers were combined to one category for significance testing;
Statistically significant (p ≤ .05); SD = standard deviation; BMI = body mass index
Symptom Severity
Approximately half of the participants (54.55%) were classified in the high psychoneurological symptom group versus low symptom group. The median number of mild to more severe symptoms reported by those in the high symptom group was three (range=2–5) while those in the low symptoms group reported a median of zero (range=0–1). The average t-scores for the low symptom group subscales were all below the PROMIS-29 average of 50 and ranged from an average of 42.6 (SD =3.72) to 48.6 (SD =6.50); the mean t-scores for the high symptom group subscales were all above 50 and ranged from 52.9 (SD =11.00) to 60.4 (SD =7.35). Three symptoms, fatigue (p<.01), anxiety (p<.01), and depression (p=.01), were also statistically different between groups, demonstrating that the symptom group stratification effectively split participants based on symptom burden. Overall, the participants reported fatigue as their most severe symptom with a mean t-score of 52.5 (SD =8.78). Of note, for the high symptom group the average reported fatigue t-score (57.2; SD=6.91) fell within the mild symptom severity range and the mean anxiety t-score (60.4, SD=7.35) was in the moderate symptom severity range. The complete symptom data is displayed in Table 2.
Table 2.
Symptom Severity by Symptom Group
| Low (N=10) | High (N=12) | Overall (N=22) | p-value | |
|---|---|---|---|---|
| Fatigue Tscore | ||||
| Mean (SD) | 46.9 (7.53) | 57.2 (6.91) | 52.5 (8.78) | .0036* |
| Median [Min, Max] | 48.6 [33.7, 53.1] | 56.1 [48.6, 69.0] | 53.1 [33.7, 69.0] | |
| Anxiety Tscore | ||||
| Mean (SD) | 42.6 (3.72) | 60.4 (7.35) | 52.3 (10.8) | .0003* |
| Median [Min, Max] | 40.3 [40.3, 48.0] | 63.4 [40.3, 67.3] | 51.9 [40.3, 67.3] | |
| Depression Tscore | ||||
| Mean (SD) | 44.0 (4.83) | 53.5 (9.22) | 49.2 (8.84) | .0113* |
| Median [Min, Max] | 41.0 [41.0, 51.8] | 54.8 [41.0, 69.4] | 49.0 [41.0, 69.4] | |
| Sleep Tscore | ||||
| Mean (SD) | 48.6 (6.50) | 54.7 (8.28) | 51.9 (7.98) | .0666 |
| Median [Min, Max] | 48.4 [41.1, 61.7] | 54.3 [41.1, 73.3] | 51.5 [41.1, 73.3] | |
| Pain Tscore | ||||
| Mean (SD) | 46.8 (6.86) | 52.9 (11.0) | 50.1 (9.68) | .1800 |
| Median [Min, Max] | 41.6 [41.6, 57.1] | 54.8 [41.6, 71.6] | 46.8 [41.6, 71.6] | |
| Number of Symptoms ≥ Mild Severity | ||||
| Mean (SD) | 0.300 (0.483) | 3.00 (1.13) | 1.77 (1.63) | |
| Median [Min, Max] | 0 [0, 1.00] | 3.00 [2.00, 5.00] | 2.00 [0, 5.00] |
Statistically significant (p ≤ .05); SD = standard deviation
Relative Abundance
The most abundant genera in this study were Veillonella, Prevotella, Streptococcus, and Rothia. For all the participants, the 15 most abundant genera accounted for over 75% of the total genus-level relative abundance. The relative abundance graph of each participant split into the high and low symptoms groups is shown in Figure 1. The overall relative abundance of the high and low symptoms groups is displayed Supplemental Figure 1.
Figure 1:
Top 15 Most Abundant Genera by Symptom Group: Genus-level relative abundance of the participants grouped by symptom cluster (i.e., high symptom cluster group or low symptom cluster group). The 15 most abundant genera are displayed in order of overall relative abundance across participants with the less abundant genera grouped into the “other” category. The participants are also ordered by relative abundance of the most abundant genera.
Alpha And Beta Diversity
A marginally statistically significant difference between symptom groups was seen for observed ASVs (p=.05) but not for Shannon diversity nor Inverse Simpson diversity after adjustment (Figure 2). No statistical differences were detected between the high and low symptom groups for the beta diversity metrics. Visualization of the beta diversity PCoA plots seen in Supplemental Figure 2 also did not highlight any meaningful differences.
Figure 2:
Alpha Diversity by Symptom Group: Alpha diversity boxplots of the three alpha diversity metrics stratified by symptom cluster group (i.e., high symptom cluster group or low symptom cluster group). InvSimpson = Inverse Simpson; Observed = Observed amplicon sequence variants
Differential Abundance
Three genera of the 71 tested genera were differentially abundant between the symptom groups, Aggregatibacter (p-adj=.01, log2 fold change=3.77, SE=.83), Lautropia (p-adj=.01, log2 fold change=2.06, SE=.47), and Prevotellaceae UCG-001 (p-adj=.04, log2 fold change=1.78, SE=.48). All three genera were more abundant in the high symptom group. No genera were statistically more abundant in the low symptom group. Of the significant genera Aggregatibacter was associated with the largest log2 fold change with an estimated change of 3.77, which corresponds to over a 13-fold higher abundance in the high symptom group compared with the low symptom group. However, while the difference between symptom groups was large, Aggregatibacter had low overall relative abundance in this study, with a mean abundance of 1.19 × 10−4 in the low-symptom group and 4.29 × 10−4 in the high symptom group. The differential abundance results for all the tested genera and the mean relative abundance for each genus split by symptom group is shown in Table 3.
Table 3:
Differential Abundance Between Symptom Groups
| Genus | log2 Fold Change* | Standard Error | p-value | p-adj value | Mean Abundance - Low | Mean Abundance - High |
|---|---|---|---|---|---|---|
| Aggregatibacter | 3.765706 | 0.834093 | .000306 | .014688** | 0.000119 | 0.000429 |
| Lautropia | 2.063214 | 0.471713 | .000414 | .014688** | 0 | 4.36E-05 |
| Prevotellaceae UCG-001 | 1.779559 | 0.484202 | .001876 | .0444** | 0 | 3.61E-05 |
| Abiotrophia | 2.28316 | 0.672861 | .003458 | .061383 | 3.61E-05 | 0.000172 |
| Shuttleworthia | 1.736191 | 0.586006 | .00872 | .12382 | 5.90E-06 | 0.000129 |
| Alloscardovia | 2.8624 | 1.073844 | .016304 | .192933 | 0 | 0.000371 |
| Escherichia-Shigella | 4.731398 | 1.849513 | .020364 | .206547 | 0.000747 | 0.00831 |
| Anaeroglobus | 1.946205 | 0.955532 | .057555 | .510802 | 7.20E-05 | 8.26E-05 |
| Campylobacter | −0.91518 | 0.53651 | .106248 | .701044 | 0.008958 | 0.004342 |
| Selenomonas | −1.97793 | 1.163814 | .107442 | .701044 | 0.006078 | 0.0019 |
| [Eubacterium] brachy group | 1.485661 | 0.893909 | .114844 | .701044 | 3.86E-05 | 0.000129 |
| Ligilactobacillus | 0.97096 | 0.590515 | .118486 | .701044 | 0 | 4.42E-05 |
| F0332 | 1.304513 | 0.880322 | .156677 | .754751 | 6.77E-05 | 0.000113 |
| Butyrivibrio | −1.24555 | 0.847594 | .159958 | .754751 | 0.000161 | 2.16E-05 |
| Fretibacterium | 1.401427 | 1.022477 | .188321 | .754751 | 8.59E-05 | 0.000231 |
| Tannerella | 1.16025 | 0.877609 | .203663 | .754751 | 0.000293 | 0.000199 |
| Dialister | 1.218823 | 0.928632 | .206799 | .754751 | 0.000104 | 0.000126 |
| Rothia | 0.753598 | 0.575432 | .207748 | .754751 | 0.055629 | 0.110255 |
| Simonsiella | −0.80508 | 0.624103 | .214333 | .754751 | 9.46E-05 | 1.13E-05 |
| Pseudoleptotrichia | −1.6009 | 1.259582 | .220854 | .754751 | 0.015951 | 0.011046 |
| Family XIII UCG-001 | −0.97015 | 0.767434 | .223236 | .754751 | 0.000135 | 7.97E-06 |
| Hoylesella | −1.70837 | 1.424505 | .24687 | .760592 | 0.029118 | 0.022478 |
| Staphylococcus | 0.806343 | 0.677354 | .250236 | .760592 | 5.24E-06 | 2.50E-05 |
| Granulicatella | 0.543466 | 0.471311 | .264828 | .760592 | 0.010322 | 0.016374 |
| Treponema | 1.388132 | 1.211659 | .267814 | .760592 | 0.0004 | 0.000463 |
| Neisseria | −1.69419 | 1.531331 | .283986 | .775501 | 0.081887 | 0.060688 |
| Fusobacterium | −0.76775 | 0.788991 | .344159 | .86658 | 0.032926 | 0.021435 |
| Mycoplasma | 0.828103 | 0.857587 | .347772 | .86658 | 0.000115 | 6.00E-05 |
| Catonella | −0.86794 | 0.933293 | .3654 | .86658 | 0.001934 | 0.000651 |
| Actinomyces | 0.875704 | 0.943169 | .36616 | .86658 | 0.0047 | 0.014391 |
| Megasphaera | −1.15242 | 1.318838 | .394397 | .8832 | 0.017455 | 0.011531 |
| Eikenella | 0.729861 | 0.888972 | .423001 | .8832 | 0.000585 | 0.00049 |
| Peptoanaerobacter | −0.89471 | 1.095165 | .425251 | .8832 | 0.00084 | 0.000207 |
| [Eubacterium] yurii group | 0.612559 | 0.756016 | .428996 | .8832 | 0.000157 | 6.59E-05 |
| Bergeyella | −0.86038 | 1.076982 | .43538 | .8832 | 0.001899 | 0.001726 |
| Oribacterium | −0.55919 | 0.755256 | .469163 | .925293 | 0.008904 | 0.007677 |
| Alloprevotella | 0.878974 | 1.34013 | .520675 | .966814 | 0.001168 | 0.001187 |
| Actinobacillus | −0.54049 | 0.82559 | .521442 | .966814 | 0.000281 | 3.33E-05 |
| Segatella | 0.486794 | 0.764017 | .532514 | .966814 | 0.011475 | 0.00942 |
| Rikenellaceae RC9 gut group | 0.390569 | 0.631857 | .544684 | .966814 | 1.62E-05 | 1.68E-05 |
| Prevotella | −0.41406 | 0.743204 | .584697 | .969522 | 0.151737 | 0.121544 |
| Streptococcus | 0.347713 | 0.628378 | .587228 | .969522 | 0.096184 | 0.145727 |
| Cardiobacterium | −0.3504 | 0.716102 | .630868 | .969522 | 0.000114 | 6.24E-05 |
| Parvimonas | 0.474654 | 1.089812 | .668652 | .969522 | 0.003731 | 0.001825 |
| Lachnoanaerobaculum | −0.24774 | 0.592706 | .681198 | .969522 | 0.005709 | 0.004343 |
| Johnsonella | 0.412588 | 1.056847 | .701093 | .969522 | 0.000974 | 0.000263 |
| Schaalia | −0.19317 | 0.497212 | .702466 | .969522 | 0.046681 | 0.036751 |
| TM7x | 0.196069 | 0.547936 | .724873 | .969522 | 1.91E-05 | 1.49E-05 |
| Solobacterium | −0.30826 | 0.870108 | .727491 | .969522 | 0.00099 | 0.000992 |
| Capnocytophaga | 0.326252 | 0.925635 | .728823 | .969522 | 0.003211 | 0.002456 |
| Slackia | 0.19595 | 0.586588 | .742427 | .969522 | 2.13E-05 | 2.40E-05 |
| Scardovia | 0.283133 | 0.889081 | .754018 | .969522 | 1.76E-05 | 3.62E-05 |
| Porphyromonas | −0.32029 | 1.024079 | .75827 | .969522 | 0.008467 | 0.008938 |
| [Eubacterium] sulci group | −0.4351 | 1.633471 | .793162 | .969522 | 0.001951 | 0.001568 |
| Alysiella | −0.21077 | 0.795128 | .794135 | .969522 | 0.000191 | 0.000135 |
| Stomatobaculum | 0.21447 | 0.837009 | .800845 | .969522 | 0.006834 | 0.003105 |
| Filifactor | 0.169717 | 0.724289 | .817534 | .969522 | 3.29E-05 | 2.70E-05 |
| Gemella | −0.23414 | 1.051878 | .826502 | .969522 | 0.014798 | 0.01764 |
| Haemophilus | −0.27615 | 1.241222 | .82659 | .969522 | 0.076357 | 0.078542 |
| Centipeda | −0.22664 | 1.078871 | .836106 | .969522 | 0.000502 | 0.000118 |
| Peptostreptococcus | 0.217352 | 1.353722 | .874333 | .969522 | 0.007241 | 0.004491 |
| Comamonas | 0.137317 | 0.868736 | .876268 | .969522 | 0.000105 | 0.00014 |
| Peptococcus | −0.15389 | 1.115266 | .891874 | .969522 | 0.000952 | 0.000251 |
| Veillonella | 0.066517 | 0.499834 | .895695 | .969522 | 0.266872 | 0.255183 |
| Bifidobacterium | −0.12373 | 0.984159 | .901425 | .969522 | 9.55E-05 | 1.81E-05 |
| Leptotrichia | −0.13391 | 1.144479 | .908228 | .969522 | 0.003065 | 0.003246 |
| Mogibacterium | 0.08612 | 0.794032 | .914901 | .969522 | 0.000893 | 0.000629 |
| Cryptobacterium | −0.05917 | 0.73656 | .936913 | .978248 | 2.82E-05 | 5.95E-06 |
| Mannheimia | 0.072988 | 1.177897 | .951314 | .978888 | 0.003814 | 0.001274 |
| Corynebacterium | −0.01677 | 0.728836 | .981908 | .98476 | 0.0006 | 0.000311 |
| Lancefieldella | −0.01593 | 0.821796 | .98476 | .98476 | 0.00505 | 0.00342 |
The low symptom group was the reference group, so a positive value represents that the genus is more common in the high symptom group than the low symptom group;
Statistically significant (p-adj ≤ .05)
DISCUSSION
This study found differences in both microbial diversity and composition between the high and low psychoneurological symptom cluster groups in our cohort of 22 individuals with breast cancer. For alpha diversity, the high symptom group had statistically lower observed ASVs; no difference in the other alpha diversity metrics were significant. Observed ASVs is a metric that measures the total number of unique amplicon sequence variants in a sample, so lower observed ASVs in the high symptom group indicates that this group had fewer unique microbial bacteria than the low symptom group. Unlike the gut microbiome where higher alpha diversity is commonly associated with better health (Manor et al., 2020), the relationship between the oral microbiome and alpha diversity is less clear as a high alpha diversity has been associated with alcohol use, elevated body mass index, and periodontitis in healthy adults (Chaturvedi et al., 2025). Thus, more research is needed to better understand both how alpha diversity of the oral cavity affects overall health as well as the relationship between the psychoneurological symptom cluster and oral diversity.
When examining the symptom groups in relation to microbial composition, three genera (Aggregatibacter, Lautropia, and Prevotellaceae UCG-001) had an elevated relative abundance in the high symptom cluster group in comparison to the low group. In relation to prior work, oral Aggregatibacter was identified to be tangentially linked with sleep disturbances, as a previous study showed an elevated relative abundance of oral Aggregatibacter in individuals with obstructive sleep apnea–hypopnea syndrome versus healthy controls (Ko et al., 2019). Yet, while obstructive sleep apnea-hypopnea syndrome can cause sleep disturbances, the etiologies of the conditions are different, so this relationship is indirect at best. Similarly, Lautropia was linked with trait anxiety in a study of pregnant women (Alex et al., 2024). However, a systematic review and meta-analysis of 20 case-control studies examining the relationship between the oral microbiome and major psychiatric conditions found that Lautropia was increased in individuals with schizophrenia; as the major psychiatric conditions of the individuals in the high symptom group are unknown, this finding could be confounded by this variable (Cao et al., 2025). However, no associations between any major psychiatric condition and Prevotellaceae UCG-001 were reported and the relationship between Aggregatibacter and anxiety state disorder was inconsistent which could indicate that these genera are not a result of the confounding major psychiatric condition in the high symptom group (Cao et al., 2025). Associations between specific bacteria and symptoms are important because they raise the possibility that specific bacteria contribute to symptom regulation, underscoring the need for further studies to determine causality and evaluate whether targeted oral bacterial interventions could influence symptom severity.
In terms of the potential mechanistic role of Aggregatibacter, Lautropia, and Prevotellaceae UCG-001 in the oral microbiome, Aggregatibacter has been studied due to its potential to promote inflammation. Specifically, a species within the Aggregatibacter genera, Aggregatibacter actinomycetemcomitans, has been shown to induce the production of both reactive oxygen species and inflammatory cytokines (Belibasakis et al., 2019). This is relevant as numerous studies have linked inflammation to the psychoneurological symptom cluster in individuals with breast cancer (Lyon et al., 2014; Starkweather et al., 2013). Additionally, it has been hypothesized that specific oral microbes may trigger an inflammatory response (Moreno et al., 2023; Rajasekaran et al., 2024) which then may place an individual at higher risk of developing persistent inflammatory-related symptoms, such as those identified in the psychoneurological symptom cluster (George et al., 2020; Liu et al., 2012). Should a causal relationship be identified between psychoneurological symptom cluster severity and specific oral bacteria, targeted interventions, including dietary approaches to modify the oral microbiome, may offer a strategy to mitigate symptom cluster burden in individuals with breast cancer. However, it should be noted that the reverse could also be occurring in which an inflammatory environment associated with the psychoneurological symptom cluster could be altering the oral microbiome and subsequently favoring the growth of specific genera. The relationship between oral Lautropia and Prevotellaceae UCG-001 and inflammation is less clear. Longitudinal research to determine directionality is needed to understand the relationship between the oral microbiome, inflammation, and the psychoneurological symptom cluster.
In addition to the microbiome findings, participants’ overall symptom subscale t-scores were close to the population average of 50; however, significant group differences were identified for fatigue, anxiety, and depression. Clinically, the high-symptom group demonstrated mean fatigue scores exceeding the mild threshold and mean anxiety scores exceeding the moderate threshold. As pretreatment anxiety and fatigue are risk factors for increased fatigue and worse quality of life during treatment (Bardwell & Ancoli-Israel, 2008; Hughes et al., 2020), proactive monitoring of symptoms is warranted to optimize symptom management for their patients. Of note, the elevated anxiety could be related to the fact that all the participants had recently received a breast cancer diagnosis and were imminently undergoing a mastectomy. It also could be confounded by the fact that most people in the high symptom group reported either a history of or current major psychiatric condition (n=10, 83.3%) so the severe symptom levels could be related to the psychiatric condition rather than the breast cancer. However, regardless of the cause, thoughtful attention to anxiety levels within this population is warranted to manage their symptom burden.
Limitations
One limitation of this study was the small sample size. Due to the complexity and sparsity of microbiome data, this study had limited power to detect differences in abundance between the high and low symptom groups. Given this small sample size, we were only able to adjust for smoking status, alcohol use, and age, so residual confounding is possible. Future studies should use a larger sample size to explore the relationship between the psychoneurological symptom cluster and the oral microbiome in individuals with breast cancer to better evaluate confounding by factors such as antibiotic use, oral care practices, cancer stage, treatment regimen, and breast cancer subtype, in addition to investigating potential interactions. Another limitation of this study was the unequal distribution of individuals with a history of or current major psychiatric conditions between the two symptom cluster groups. This resulted in an inability to evaluate whether the symptom cluster or psychiatric conditions were driving the microbiome associations due to the almost complete separation between symptoms groups. Many major psychiatric conditions, such as depression and anxiety disorders, can affect psychoneurological symptom severity and are associated with altered oral microbial composition (Simpson et al., 2020; Wingfield et al., 2021). Future studies should either exclude individuals with major psychiatric conditions or design the study to allow for the control of this potential confounder. However, given that two of the three bacteria associated with the high-symptom group were not strongly linked to psychiatric conditions in a recent systematic review and meta-analysis, the association observed in this study is more consistent with symptom severity than with psychiatric illness.
A final limitation was the cross-sectional nature of the study which eliminated the ability to determine the directionality of the relationship between the oral microbiome and the psychoneurological symptom cluster. While the oral microbiome is posited to affect this symptom cluster, the psychoneurological symptom cluster and its associated attributes, such as inflammation, could also be modulating the oral microbiome. Thus, longitudinal studies are needed to elucidate the direction of the relationship between the oral microbiome and the psychoneurological symptom cluster in people with breast cancer. Regardless of these limitations, this study provides important preliminary evidence on the potential link between the oral microbiome and the psychoneurological symptom cluster in people with breast cancer.
CONCLUSIONS
Many people with breast cancer experience severe psychoneurological symptoms but interventions to manage these symptoms are limited as the etiology of the psychoneurological symptom cluster is poorly understood. Preliminary evidence has suggested that the oral microbiome is associated with the psychoneurological symptom cluster in both healthy and chronically ill populations. Our study found that people with breast cancer in the high symptom group had statistically lower observed ASVs than the low symptom group; no differences in beta diversity were found. The high symptom cluster group also had higher relative abundance of Aggregatibacter, Lautropia, and Prevotellaceae UCG-001. As Aggregatibacter has pro-inflammatory potential, and inflammation is known to influence psychoneurological symptom cluster severity, further investigation into the relationship between the oral microbiome and the psychoneurological symptom cluster in people with breast cancer is warranted. A better understanding of whether specific oral genera are causally linked with the psychoneurological symptom cluster severity could inform symptom management interventions to improve the symptom experience for individuals with breast cancer through modulation of the oral microbiome.
Supplementary Material
Supplemental Figure 1: Overall Relative Abundance by Symptom Group: Overall genera level relative abundance grouped by symptom cluster (i.e., high symptom cluster group or low symptom cluster group). The 15 most abundant genera are displayed in order of overall relative abundance with the less abundant genera grouped into the “other” category.
Supplemental Figure 2: Principal Coordinates Analysis: Principal coordinates analysis plots of the samples for the two beta diversity metrics. The points are displayed in shapes based on symptom cluster group to indicate whether it is part of the high symptom cluster group or low symptom cluster group. PC = Principal component; PCoA = Principal coordinates analysis
Acknowledgement:
This research was supported [in part] by the Intramural Research Program of the National Institutes of Health (NIH) and Federal Funds from the National Cancer Institute, National Institutes of Health, under Contract No. 75N91019D00024. This work was also supported by the National Institutes of Health through the National Cancer Institute under award number F31CA294598. The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. Government. The contributions of the NIH author(s) are considered Works of the United States Government. The findings and conclusions presented in this paper are those of the author(s) and do not necessarily reflect the views of the NIH or the U.S. Department of Health and Human Services. This research was also partially funded by the Triangle Center for Evolutionary Medicine and the Duke Microbiome Center. We would like to acknowledge the 1KP research team and participants who participated in this research and made this work possible. We would also like to thank the Duke University School of Medicine for the use of the Microbiome Core Facility, which provided microbiome sequencing services.
Footnotes
Conflict Of Interest: We have no known conflicts of interest to disclose
Ethical Conduct of Research: This study was approved by the Duke University Health System Institutional Review Board.
Contributor Information
Julia Slack, PhD Candidate at Duke University School of Nursing, Durham, NC..
Emily Vogtmann, Earl Stadtman Investigator at the National Cancer Institute, Bethesda, MD..
Sharron Docherty, Associate Professor and the Vice Dean of Research at the Duke University School of Nursing, Durham, NC..
Tara A. Albrecht, Walter Schroeder Endowed Nurse Research Scientist at Advocate Health, Milwaukee, WI. At the time this research was completed, she was Assistant Professor at the Duke University School of Nursing, Durham, NC..
REFERENCES
- Alex AM, Levendosky AA, Bogat GA, Muzik M, Nuttall AK, Knickmeyer RC, & Lonstein JS. (2024). Stress and mental health symptoms in early pregnancy are associated with the oral microbiome. BMJ Mental Health, 27(1), e301100. 10.1136/bmjment-2024-301100 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Apprill A, McNally S, Parsons R, & Weber L. (2015). Minor revision to V4 region SSU rRNA 806R gene primer greatly increases detection of SAR11 bacterioplankton. Aquatic Microbial Ecology, 75(2), 129–137. 10.3354/ame01753 [DOI] [Google Scholar]
- Bardwell WA, & Ancoli-Israel S. (2008). Breast cancer and fatigue. Sleep Medicine Clinics, 3(1), 61–71. 10.1016/j.jsmc.2007.10.011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Belibasakis GN, Maula T, Bao K, Lindholm M, Bostanci N, Oscarsson J, Ihalin R, & Johansson A. (2019). Virulence and pathogenicity properties of aggregatibacter actinomycetemcomitans. Pathogens, 8(4), 222. 10.3390/pathogens8040222 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Callahan BJ, McMurdie PJ, Rosen MJ, Han AW, Johnson AJA, & Holmes SP. (2016). DADA2: High-resolution sample inference from illumina amplicon data. Nature Methods, 13(7), 581–583. 10.1038/nmeth.3869 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Cao D, Yang J, He Y, Zheng X, Li Y, Chen Y, & Tu Y. (2025). Altered oral microbiome composition in mental disorders: A systematic review and meta-analysis. Journal of Oral Microbiology, 17(1), 2541828. 10.1080/20002297.2025.2541828 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Caporaso JG, Lauber CL, Walters WA, Berg-Lyons D, Huntley J, Fierer N, Owens SM, Betley J, Fraser L, Bauer M, Gormley N, Gilbert JA, Smith G, & Knight R. (2012). Ultra-high-throughput microbial community analysis on the Illumina HiSeq and MiSeq platforms. The ISME Journal, 6(8), Article 8. 10.1038/ismej.2012.8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Carbone E, Fabrizi E, Rivabene R, Cappella M, Fortini P, Conti L, Locuratolo N, Lorenzini P, Lacorte E, & Piscopo P. (2025). Human oral microbiome in aging: A systematic review. Mechanisms of Ageing and Development, 226, 112080. 10.1016/j.mad.2025.112080 [DOI] [PubMed] [Google Scholar]
- Cella D, Choi SW, Condon DM, Schalet B, Hays RD, Rothrock NE, Yount S, Cook KF, Gershon RC, Amtmann D, DeWalt DA, Pilkonis PA, Stone AA, Weinfurt K, & Reeve BB. (2019). PROMIS® adult health profiles: Efficient short-form measures of seven health domains. Value in Health: The Journal of the International Society for Pharmacoeconomics and Outcomes Research, 22(5), 537–544. 10.1016/j.jval.2019.02.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chattopadhyay S, Malayil L, Chopyk J, Smyth E, Kulkarni P, Raspanti G, Thomas SB, Sapkota A, Mongodin EF, & Sapkota AR. (2024). Oral microbiome dysbiosis among cigarette smokers and smokeless tobacco users compared to non-users. Scientific Reports, 14(1), 10394. 10.1038/s41598-024-60730-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chaturvedi AK, Vogtmann E, Shi J, Yano Y, Blaser MJ, Bokulich NA, Caporaso JG, Gillison ML, Graubard BI, Hua X, Hullings AG, Kahle L, Knight R, Li S, McLean J, Purandare V, Wan Y, Freedman ND, & Abnet CC. (2025). Oral microbiome profile of the US population. JAMA Network Open, 8(5), e258283. 10.1001/jamanetworkopen.2025.8283 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Davis NM, Proctor DM, Holmes SP, Relman DA, & Callahan BJ. (2018). Simple statistical identification and removal of contaminant sequences in marker-gene and metagenomics data. Microbiome, 6(1), 226. 10.1186/s40168-018-0605-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Doong S-H, Dhruva A, Dunn LB, West C, Paul SM, Cooper BA, Elboim C, Abrams G, Merriman JD, Langford DJ, Leutwyler H, Baggott C, Kober K, Aouizerat BE, & Miaskowski C. (2015). Associations between cytokine genes and a symptom cluster of pain, fatigue, sleep disturbance, and depression in patients prior to breast cancer surgery. Biological Research for Nursing, 17(3), 237–247. 10.1177/1099800414550394 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Fan X, Peters BA, Jacobs EJ, Gapstur SM, Purdue MP, Freedman ND, Alekseyenko AV, Wu J, Yang L, Pei Z, Hayes RB, & Ahn J. (2018). Drinking alcohol is associated with variation in the human oral microbiome in a large study of American adults. Microbiome, 6(1), 59. 10.1186/s40168-018-0448-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- George MA, Lustberg MB, & Orchard TS. (2020). Psychoneurological symptom cluster in breast cancer: The role of inflammation and diet. Breast Cancer Research and Treatment, 184(1), 1–9. 10.1007/s10549-020-05808-x [DOI] [PubMed] [Google Scholar]
- Grayson S, Sereika S, Conley Y, Bender C, Carr K, & Wesmiller S. (2023). Psychoneurological symptom cluster trajectories in individuals treated for early-stage breast cancer. Supportive Care in Cancer, 31(10), 563. 10.1007/s00520-023-08018-5 [DOI] [PubMed] [Google Scholar]
- Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, & Conde JG. (2009). Research electronic data capture (REDCap)—A metadata-driven methodology and workflow process for providing translational research informatics support. Journal of Biomedical Informatics, 42(2), 377–381. 10.1016/j.jbi.2008.08.010 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hays RD, Spritzer KL, Schalet BD, & Cella D. (2018). Promis®−29 v2.0 profile physical and mental health summary scores. Quality of Life Research : An International Journal of Quality of Life Aspects of Treatment, Care and Rehabilitation, 27(7), 1885–1891. 10.1007/s11136-018-1842-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hughes A, Suleman S, Rimes KA, Marsden J, & Chalder T. (2020). Cancer-related fatigue and functional impairment—Towards an understanding of cognitive and behavioural factors. Journal of Psychosomatic Research, 134, 110127. 10.1016/j.jpsychores.2020.110127 [DOI] [PubMed] [Google Scholar]
- Kers JG, & Saccenti E. (2022). The power of microbiome studies: Some considerations on which alpha and beta metrics to use and how to report results. Frontiers in Microbiology, 12, 796025. 10.3389/fmicb.2021.796025 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Khajoei R, Ilkhani M, Azadeh P, Zohari Anboohi S, & Heshmati Nabavi F. (2023). Breast cancer survivors–supportive care needs: Systematic review. BMJ Supportive & Palliative Care, 13(2), 143–153. 10.1136/spcare-2022-003931 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kidwell KM, Harte SE, Hayes DF, Storniolo AM, Carpenter J, Flockhart DA, Stearns V, Clauw DJ, Williams DA, & Henry NL. (2014). Patient-reported symptoms and discontinuation of adjuvant aromatase inhibitor therapy. Cancer, 120(16), 2403–2411. 10.1002/cncr.28756 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ko C-Y, Hu A-K, Chou D, Huang L-M, Su H-Z, Yan F-R, Zhang X-B, Zhang H-P, & Zeng Y-M. (2019). Analysis of oral microbiota in patients with obstructive sleep apnea-associated hypertension. Hypertension Research, 42(11), 1692–1700. 10.1038/s41440-019-0260-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Li G. (2021). A Review of the Research on Symptom Clusters in Cancer Survivors. Open Journal of Nursing, 11(6), 423–434. 10.4236/ojn.2021.116036 [DOI] [Google Scholar]
- Liu L, Mills PJ, Rissling M, Fiorentino L, Natarajan L, Dimsdale JE, Sadler GR, Parker BA, & Ancoli-Israel S. (2012). Fatigue and sleep quality are associated with changes in inflammatory markers in breast cancer patients undergoing chemotherapy. Brain, Behavior, and Immunity, Aging, Brain, Behavior, and Immunity, 26(5), 706–713. 10.1016/j.bbi.2012.02.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lupo GFD, Rocchetti G, Lucini L, Lorusso L, Manara E, Bertelli M, Puglisi E, & Capelli E. (2021). Potential role of microbiome in chronic fatigue syndrome/myalgic encephalomyelits (CFS/ME). Scientific Reports, 11(1), 7043. 10.1038/s41598-021-86425-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lyon D, Elmore L, Aboalela N, Merrill-Schools J, McCain N, Starkweather A, Elswick RK, & Jackson-Cook C. (2014). Potential epigenetic mechanism(s) associated with the persistence of psychoneurological symptoms in women receiving chemotherapy for breast cancer: A hypothesis. Biological Research for Nursing, 16(2), 160–174. 10.1177/1099800413483545 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Manor O, Dai CL, Kornilov SA, Smith B, Price ND, Lovejoy JC, Gibbons SM, & Magis AT. (2020). Health and disease markers correlate with gut microbiome composition across thousands of people. Nature Communications, 11(1), Article 1. 10.1038/s41467-020-18871-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- McMurdie PJ, & Holmes S. (2013). phyloseq: An R package for reproducible interactive analysis and graphics of microbiome census data. PLOS ONE, 8(4), e61217. 10.1371/journal.pone.0061217 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Miaskowski C, Barsevick A, Berger A, Casagrande R, Grady PA, Jacobsen P, Kutner J, Patrick D, Zimmerman L, Xiao C, Matocha M, & Marden S. (2017). Advancing symptom science through symptom cluster research: Expert panel proceedings and recommendations. JNCI: Journal of the National Cancer Institute, 109(4), djw253. 10.1093/jnci/djw253 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Moreno CM, Boeree E, Freitas CMT, & Weber KS. (2023). Immunomodulatory role of oral microbiota in inflammatory diseases and allergic conditions. Frontiers in Allergy, 4, 1067483. 10.3389/falgy.2023.1067483 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Peng X, Han N, Gong Y, He L, Xu Y, Xiao D, Zhang T, Qiang Y, Li X, Zhang W, & Zhang J. (2025). Characterization of the salivary microbiome in healthy individuals under fatigue status. Frontiers in Cellular and Infection Microbiology, 15, 1506723. 10.3389/fcimb.2025.1506723 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Qiu X, Xu T, Huang Y, Wei C, Wang L, & Wu B. (2025). Relationship between depression and oral microbiome diversity: Analysis of NHANES data (2009–2012). BMC Oral Health, 25(1), 914. 10.1186/s12903-025-06274-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- Rajasekaran JJ, Krishnamurthy HK, Bosco J, Jayaraman V, Krishna K, Wang T, & Bei K. (2024). Oral microbiome: A review of its impact on oral and systemic health. Microorganisms, 12(9), 1797. 10.3390/microorganisms12091797 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Richardson LC, Wang W, Hartzema AG, & Wagner S. (2007). The role of health-related quality of life in early discontinuation of chemotherapy for breast cancer. The Breast Journal, 13(6), 581–587. 10.1111/j.1524-4741.2007.00512.x [DOI] [PubMed] [Google Scholar]
- Simpson CA, Adler C, du Plessis MR, Landau ER, Dashper SG, Reynolds EC, Schwartz OS, & Simmons JG. (2020). Oral microbiome composition, but not diversity, is associated with adolescent anxiety and depression symptoms. Physiology & Behavior, 226, 113126. 10.1016/j.physbeh.2020.113126 [DOI] [PubMed] [Google Scholar]
- Starkweather AR, Lyon DE, Elswick RK, Montpetit AJ, Conley Y, & McCain NL. (2013). A conceptual model of psychoneurological symptom cluster variation in women with breast cancer: Bringing nursing research to personalized medicine. Current Pharmacogenomics and Personalized Medicine, 11(3), 224–230. 10.2174/18756921113119990004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Tometich D, Small BJ, Carroll JE, Zhai W, Luta G, Zhou X, Kobayashi LC, Ahles T, Saykin AJ, Clapp JD, Jim HSL, Jacobsen PB, Hurria A, Graham D, McDonald BC, Denduluri N, Extermann M, Isaacs C, Dilawari A, … Mandelblatt JS. (2019). Pre-treatment psychoneurological symptoms and their association with longitudinal cognitive function and quality of life in older breast cancer survivors. Journal of Pain and Symptom Management, 57(3), 596–606. 10.1016/j.jpainsymman.2018.11.015 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wingfield B, Lapsley C, McDowell A, Miliotis G, McLafferty M, O’Neill SM, Coleman S, McGinnity TM, Bjourson AJ, & Murray EK. (2021). Variations in the oral microbiome are associated with depression in young adults. Scientific Reports, 11(1), 15009. 10.1038/s41598-021-94498-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Zhou H, He K, Chen J, & Zhang X. (2022). LinDA: Linear models for differential abundance analysis of microbiome compositional data. Genome Biology, 23(1), 95. 10.1186/s13059-022-02655-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplemental Figure 1: Overall Relative Abundance by Symptom Group: Overall genera level relative abundance grouped by symptom cluster (i.e., high symptom cluster group or low symptom cluster group). The 15 most abundant genera are displayed in order of overall relative abundance with the less abundant genera grouped into the “other” category.
Supplemental Figure 2: Principal Coordinates Analysis: Principal coordinates analysis plots of the samples for the two beta diversity metrics. The points are displayed in shapes based on symptom cluster group to indicate whether it is part of the high symptom cluster group or low symptom cluster group. PC = Principal component; PCoA = Principal coordinates analysis