Published in ASCO (American Society of Clinical Oncology) and MASCC (Multinational Association of Supportive Care in Cancer) conference proceedings.
Association of social determinant of health (SDOH) with clinical outcomes in patients with cancer: opioid use, emergency admission, anxiety, depression, and length of hospital stays. Satheesh Kumar Poolakkad Sankaran, Minu Ponnamma Mohan, Alam Naheed, Joel Epstein, Robert Pili. Association of social determinant of health (SDOH) with clinical outcomes in patients with cancer: Opioid use, emergency admission, anxiety, depression, and length of hospital stays. JCO 42, e23072-e23072(2024). DOI:10.1200/JCO.2024.42.16_suppl.e23072
As cancer patients are overburdened by disease, treatment, access to care, and other demographic factors, the obstacles become increasingly significant with SDOH. Patients hospitalized in the United States in 2017 for the treatment of prostate (PC), breast (BC), and lung (LC) cancer were evaluated in order to further characterize and contrast variations between SDOH and clinical outcomes.
We utilized generalized linear models to assess the association between homelessness and problem related living alone (PrbLA) and clinical outcomes–opioid abuse/long-term use, depression, and anxiety, and burden of illness (length of stay), while controlling for patients’ characteristics.
In the study among PC there were 425 homeless, 355 PrbLA; BC, 320 homeless and 225 PrbLA; LC, 1345 homeless and 820 PrbLA. In the adjusted analysis, among PC homelessness were significantly associated with anxiety and depression (5.15, 95%CI: 3.17-8.35); emergency admission, (5.57; 95%CI: 2.63-11.82); opioid abuse (3.75 ,95%CI: 1.27-11.13) and longer LOS, unadjusted (2.49; 95%CI: 1.58-3.92), adjusted analysis (1.96; 95%CI: 1.03-3.74). These findings were comparable in the BC and LC cohorts. Further in the adjusted analysis, PC patients with PrbLA were associated with higher combined anxiety and depression, adjusted (2.7, 95%CI: 1.64-4.46); emergency admission, adjusted (5.86; 95%CI: 2.66-12.91); opioid long-term use, adjusted (3.74, 95%CI: 1.02-6.88) and longer LOS, adjusted analysis (1.44; 95%CI: 1.11-1.88). These findings were comparable in the BC cohort. In the LC cohort when stratified to females, the PrbLA was associated with longer LOS (1.29; 95%CI: 1.02-1.63). In all cancer cohorts, when compared to Whites, SDOH was associated with poor outcomes; Blacks had a higher burden of illness. And when compared to Whites, Hispanics, Asians, and Others were associated with lesser opioid abuse among all cancer cohorts with homelessness.
According to this study, PC, BC, and LC patients with poor SDOH presented with worst clinical outcomes, including higher opioid abuse, emergency admission, anxiety and depression, and longer LOS in the hospital. The findings of this study highlight a vacuum in the literature, and the recommendations stress the value of social support for cancer patients in achieving a better prognosis and quality of life.
Racial and income characteristics and the burden of illness among metastatic prostate cancer patients. Satheesh kumar Poolakkad Sankaran, Joel Brian Epstein, Eric Adjei Boakye, Naheed Alam, and Roberto Pili Racial and income characteristics and the burden of illness among metastatic prostate cancer patients. JCO 41, e17064-e17064(2023). DOI:10.1200/JCO.2023.41.16_suppl.e17064
The incidence of metastatic prostate cancer has increased dramatically between 2010 and 2017 in the United States. Since the racial and ethnic mix of the population of the United States (U.S.) have changed between the years 2010 and 2020–it is extremely important to understand its effects on prostate cancer outcomes. Therefore, we investigated the burden of illness across racial groups and income levels among patients diagnosed with metastatic prostate cancer.
This study utilized the U.S. 2017 National Inpatient Sample database to identify hospitalizations with a primary diagnosis of prostate cancers. We evaluated associations between race and income levels, and the burden of illness (total charges and length of stay (LOS)) adjusting for patient and clinical factors with multivariable generalized linear models.
A total of 63,880 metastatic prostate cancer patients were included in the study, and the median [IQR] age was 74 [66, 82]. There were 65.6% Whites, 20.6% Blacks, 7.7% Hispanics, and 6.1% Asians and Others. Blacks had a higher prevalence in the age group 45-55 (34.2%), when compared to 55-64 (29.8%), or > 65 (17.6%) age group. There were only 10% Blacks and 4.9% Hispanics living in high-income neighborhoods, whereas this was 75.5% for Whites. In the adjusted analysis, compared to the Whites, Hispanics (1.43; 95%CI-1.29- 1.57), and Asians and Others (1.39; 95%CI- 1.25- 1.56) had a higher total charge; and compared to Whites, Blacks had a higher LOS (1.28; 95%CI-1.15-1.44). Compared to high-income neighborhoods, those living in the 26th to 50th percentile had a lower total charge (0.86; 95%CI- 0.78-0.94). Compared to the > 65 years age group, the 44-55 age group had a lower total charge (0.73; 95%CI-0.61-0.89). Compared to patients with Medicare, those with Private Insurance (0.87; 95%CI- 0.80-0.94) and Self-pay and Others (0.73; 95C%CI-0.62–0.86) had a lower total charge. And when compared to non-elective admissions, elective admissions had a higher total charge (1.44; 95%CI-1.29- 1.59).
Hispanics and Asians and Others had a higher burden of illness, and Blacks had a higher LOS. Patients who reside in lower income neighborhoods had lower total charges. Understanding prostate cancer outcomes by race and income is crucial to designing strategies to improve access to care.
The integration of big data analytics in examining the effects of mental illness screening on mortality and burden of illness among patients with cancer. Minu Mohan, Joel Brian Epstein, Naheed Alam, Shaniza Haniff, Roberto Pili, and Satheesh Kumar Poolakkad Sankaran. The integration of big data analytics in examining the effects of mental illness screening on mortality and burden of illness among patients with cancer. JCO 42, e23083-e23083(2024). DOI:10.1200/JCO.2024.42.16_suppl.e23083
Cancer patients receiving cancer care in US hospitals undergo notable psychological suffering, characterized by manifestations of multiple psychiatric disorders, and further doubling of healthcare spending each year. The psychological burden of illness frequently accompanies the physical obstacles of the disease, hence increasing the intricacy of cancer care. An analysis was conducted on cancer patients admitted to hospitals in the US in 2017 for the treatment of prostate (PC), cancers of the lip, oral cavity, and pharynx (CLOP), lung (LC), and leukemia. The aim was to better understand and compare the differences in mental illness screening (MIS) and its impact on these patients.
We employed generalized linear models to examine the association between mental illness screening and outcomes––Burden of illness (BOI), such as cost, total charge, length of stay (LOS), and in-hospital mortality after controlling for clinical and patients' variables.
The study found that among 224,540 LC, there were 178,470 cancer patients who underwent MIS. This was 151,285 vs. 58,125 among PC; among CLOP, this was 35,770 vs. 18,495; for leukemia this was 53,019 vs.18,759. In the adjusted analysis, MIS among LC cohort was associated with lower length of stay—coefficient, 0.96 (95% CI, 0.92-0.98), lower total charges, (0.94, 0.92-0.96), and lower mortality (aOR, 0.77, 95% CI, 0.73-0.82). In the LC cohort, compared to Whites, Blacks (1.11, 1.05-1.17), and Others (Native Americans, Asians and Others (1.1, 1.02-1.20) had longer LOS. And compared to high-income neighborhoods, low-income neighborhoods showed higher LOS (1.09, 1.03-1.153), and compared to “Central” counties of metro areas of ≥ 1 million population, Not metropolitan or micropolitan counties had a lower LOS (0.86 (0.80-0.92). Compared to Medicare beneficiaries, Private insurance holders had a lower total charge (0.84,0.81-0.88). Females when compared to males had a lower total charge (0.96, 0.94-0.98), and when compared to Whites, Hispanics (1.35, 1.25-1.45), and Others (1.28, 1.19-1.37) had a higher total-charges. And compared to Whites, Others had a higher in-hospital mortality (aOR, 1.16, 95% CI, 1.03-1.31). Further in the LC cohort, the analysis was stratified to surgical and non-surgical cohorts. Compared to LC, none of the other solid or liquid cancer cohorts, MIS showed no association of BOI or mortality.
Unlike, other solid and liquid cancers, LC cohort had a beneficial association of MIS and BOI and in-hospital mortality. In order to provide comprehensive care for all cancer patients, it is beneficial to incorporate MIS as a routine. This study warrants valuable insights for future research on the effects and execution of regular MIS among individuals with cancer.
Employing machine learning (ML) and explainable artificial intelligence (XAI) to predict and explain suicidal ideation among patients with prostate cancer. Satheesh Kumar Poolakkad Sankaran, Minu Mohan, Joel Brian Epstein, and Roberto Pili. Employing machine learning (ML) and explainable artificial intelligence (XAI) to predict and explain suicidal ideation among patients with prostate cancer. JCO 42, e23086-e23086(2024). DOI:10.1200/JCO.2024.42.16_suppl.e23086
It is anticipated that the implementation of machine learning (ML) strategies will have an impact on the clinical decision-making process in the future. The aim of this study was to compare the accuracy of various different machine learning algorithms in predicting suicidal ideation in patients who were receiving treatment for prostate cancer (PC). Further to predict suicidal ideation and explain burden of illness (BOI) by utilizing ML algorithms and XAI based interpretability.
We analyzed the United States 2017 National Inpatient Sample database for patients who were hospitalized for PC using both linear and non-linear ML methods. Using techniques such as forward selection (FS) and backward elimination (BE), Random Forest (RF), decision trees, Multivariate Adaptive Regression Splines, and Gradient Boosting Machine (GBM), we determined subsets and features. We used linear and non-linear MLs-- Lasso, Ridge, RF, and Neural Networks (NN). A partitioning of 70% and 30% was performed on the training and test datasets. The performance of the model was tested using discrimination (C-statistics), Receiver-Operating Characteristics (ROC) curve, and Hosmer-Lemeshow tests on both the training data and the test data. A number of XAI approaches, including permutation significance, global surrogate marker, feature interpretation, interactivity, and local interpretability, were also utilized to examine the BOI (Length of stay (LOS)) among PC cancer cohort with suicide ideation.
We identified 680 patients with suicidal ideation among 208,730 PC patients, The most important variable derived through FS and BE were depression, drug abuse, psychiatric disorder, alcohol dependence, race, Medicaid beneficiaries, cardiac arrythmias, metastasis, weight loss, congestive heart failure, and un-complicated hypertension; further confirmed through the GBM, and other methods. There was a successful documentation of demographic, socioeconomic, and clinical characteristics through the use of feature selection approaches. Linear models, Lasso, and Ridge demonstrated an excellent area under the ROC for all cohorts (>0.9, both in train and test datasets). On the other hand, tree-based models, such as RF, performed poor (AUC 0.72 for train and 0.66 for test) (Figure 1). According to the findings of the XAI methodologies, showed that among PC suicide ideation, higher BOI was associated with patients staying in low-income neighborhoods, Whites, Medicaid beneficiaries, and those with weight loss (Figure 2, ALE local interpretability plot).
We showcase the increasing capability of predictive analytics, particularly XAI, in predicting suicidal ideation. The combination of the vast amount of data generated by the digitalization of healthcare systems with the advanced processing power of modern servers enables the development of machine learning-based care pathways, which has the potential to revolutionize cancer clinical decision-making.
Cancer patients with suicide ideation are associated with higher burden of illness, depressive disorders, and alcohol and substance abuse disorders. (GU cancers). Abstracts for MASCC/AFSOS/ISOO Annual Meeting 2024. Support Care Cancer 32 (Suppl 1), 434 (2024). https://doi.org/10.1007/s00520-024-08541-z
Authors: Minu Ponnamma Mohan, DDS, MDS, MPH1; Roberto Pili, MD2; Joel B Epstein, DMD, MSD3; Venu Gopalakrishnan, MD4; Sreelakshmi Panginikkode, MD5; Susan Eichhorn, MD6; Naheed Alam, MD7; Poolakkad S Satheeshkumar, DDS, MDS, MSc, MMSc8
Suicide tendency is a serious concern among cancer patients; nevertheless, subsequent investigations have not significantly explored the impact of a cancer diagnosis on the psychological health of at-risk patients. Furthermore, it is critical to investigate the potential impact of treatment accessibility on the risk of increased burden of illness (BOI), disparities, and associated mental illness among patients who have suicidal ideation (SI). We assessed the cancer patients with SI who were admitted to the hospital for treatment of genitourinary (GU) malignancies.
For this research we used National Inpatient Sample database. We used generalized linear models to evaluate the association of SI and the outcome of BOI, and further we evaluated the association between SI and depressive disorders and alcohol and substance abuse disorders, while accounting for patient and clinical characteristics.
Among GU bladder cancers (80,685) there were 235 SI patients, and prostate cancers (PC) (209,410) with 680 SI. PC patients with SI had a longer hospital length of stay (coefficient, 3.14; 95%CI, 2.43-4.79), higher depressive disorders (aOR, 19.45, 95%CI, 13.53 27.96), higher alcohol related disorders (aOR, 8.81, 95%CI, 5.61-13.83), and substance abuse (aOR, 5.17, 95%CI, 3.61-7.39). Bladder ca patients with SI had had a longer hospital length of stay (coefficient, 2.38; 95%CI, 1.62-3.52), higher depressive disorders (aOR, 8.32, 95%CI, 4.64-14.91), higher alcohol related disorders (aOR, 7.78, 95%CI, 3.19-18.87), and substance abuse (aOR, 6.17, 95%CI, 3.27-11.65).
This research study specifically examines suicides associated with cancer in the United States and indicates areas where additional research is required to address psychosocial needs. Findings enhances the existing knowledge utilized for making well-informed decisions regarding measures for avoiding suicide, providing psychological support, reducing alcohol and substance addiction. Continuous research on identifying and quantifying the extent of suicide risk among cancer patients is highly important.
Hospitalized cancer patients with opioid-dependency are associated with increased burden of illness, depressive disorders, and septicemia (GU cancers) Abstracts for MASCC/AFSOS/ISOO Annual Meeting 2024. Support Care Cancer 32 (Suppl 1), 434 (2024). https://doi.org/10.1007/s00520-024-08541-z
Authors: Minu Ponnamma Mohan, DDS, MDS, MPH1; Roberto Pili, MD2; Joel B Epstein, DMD, MSD3; Venu Gopalakrishnan, MD4; Sreelakshmi Panginikkode, MD5; Susan Eichhorn, MD6; Naheed Alam, MD7; Poolakkad S Satheeshkumar, DDS, MDS, MSc, MMSc
From the early phases of cancer pain treatment to palliative care in the later stages, opioids are quite effective. Still, a lot is unknown about opioid dependence (OD) and its impact on cancer patients. Both biological and psychological variables contribute to the development of OD, making it a multi-faceted condition. In addition to coping with cancer's own symptoms, mental health concerns and infectious consequences are additional obstacles that cancer patients must overcome, leading to out-of-pocket costs and barriers to access to care.
We utilized national inpatient database for this research. While accounting for patient and clinical characteristics, we used generalized linear models to evaluate the association of OD and the outcome of BOI, and further we evaluated the association between OD and mental illness screening, anxiety and depression disorders, and septicemias.
Among GU renal cell cancers (RCC) (88105) there were 695 OD patients, and prostate cancers (PC) (209,410) with 1115 OD. PC patients with OD were associated with longer hospital length of stay (Coefficient, 1.71; 95%CI, 1.28-2.29), lesser mental illness screening (aOR: 0.67, 95%CI, 0.47-0.96), higher anxiety disorders (aOR,2.47, 95%CI, 1.7-3.59) depressive disorders (aOR, 2.13, 95%CI, 1.46-3.01), and septicemia (aOR, 2.22, 95%CI, 1.54-3.21). RCC patients with OD were associated with a longer hospital length of stay (Coefficient, 1.62; 95%CI, 1.09-2.43), higher anxiety disorders (aOR,2.06, 1.34-3.16), depressive disorders (aOR, 1.93, 95%CI, 1.24-2.99), and higher septicemia (aOR, 2.66, 95%CI, 1.59-4.46).
The availability of opioids is crucial in order to provide high-quality care for this specific patient population. However, In the exceptional clinical situation where there is a high risk of mental illness, there must be specific criteria for controlling cancer pain that enable the safe use of opioids considering the genetic and psychosocial components of OD.
Receiving immunotherapy for the treatment of advanced renal cell carcinoma was associated with higher burden of illness, coagulopathy, and cardiac arrhythmia. (GU Cancers) Abstracts for MASCC/AFSOS/ISOO Annual Meeting 2024. Support Care Cancer 32 (Suppl 1), 434 (2024). https://doi.org/10.1007/s00520-024-08541-z
Authors: Susan Eichhorn, MD, MPH1; Roberto Pili, MD2; Minu Ponnamma Mohan, MDS, MPH3; Stephen Sonis, DMD, DMSc4; Poolakkad S Satheeshkumar, DDS, MDS, MSc, MMSc
In recent years, immunotherapy (IT) has emerged as a fundamental treatment for metastatic renal cell carcinoma (mRCC). Nevertheless, the adverse events and expenses associated with IT have not been thoroughly delineated. The aim of our study is to measure association between the utilization of IT in patients with mRCC and the overall costs, coagulopathy, and arrhythmia.
A cross-sectional analysis was conducted utilizing the National Inpatient Samples database to identify and examine patients who were hospitalized with mRCC. The study examined association between the utilization of IT in mRCC and the burden of illness (BOI)—total charges, and length of stay (LOS), coagulopathy, and arrhythmia through generalized linear models (glm).
We examined 230 of the 28,535 patients with mRCC who were receiving IT. Patients receiving IT had mean [SD] age of 54.50 [8.67] vs. 65.80 [12.18] non-IT. White patients made up 81.0% of the IT recipients. 82.6% of IT beneficiaries were non-Medicaid or non-Medicare. IT patients' mean total charges were $260,905 [134,956.18] vs. $72,343 [$91,457.79] for non-IT patients, with a difference in LOS of stay of 4.98 [1.94] versus 6.28 [6.67] for non-IT patients. IT-treated mRCC patients reported a coagulopathy rate of 26.1% and an arrhythmia rate of 37.0%. In the adjusted glm, after controlling for variables, the use of IT was associated with higher total charges (7.67; 95% CI: 4.86 – 12.09). Further, coagulopathy (aOR = 5.61; 95% CI: 2.40 – 13.14) and arrhythmia (aOR = 4.34; 95% CI: 2.20 – 8.55) was associated with IT.
mRCC patients who received immunotherapy had a higher likelihood of coagulopathies and arrhythmia which impacted total charges. While IT has played a vital role in treating mRCC in recent years, this is the first instance where real-world evidence on adverse events and expenses associated with IT is being examined. The findings may have substantial implications for new supportive care in this population.
Toxicity Cluster: Complexity Reduction and Approximation Through Non-linear Principal Component Analysis With Multivariate Analysis with Optimal scaling. Abstracts for MASCC/AFSOS/ISOO Annual Meeting 2024. Support Care Cancer 32 (Suppl 1), 434 (2024). https://doi.org/10.1007/s00520-024-08541-z
Satheesh kumar Poolakkad Sankaran, DDS, MDS, Pg Diploma, ICRT, MSc, MMSc; Minu Ponnamma Mohan, DDS, MDS, MPH; Stephen Sonis, DMD, DMSc; Sreelakshmi Panginikkode, MD; Venu Gopalakrishnan, MD; Roberto Pili, MD.
Cancer treatment-induced toxicities (CIT) are a comprehensive indicator of health outcome status among cancer patients and are valuable in informing treatment options for complex clusters. This study sought to determine the appropriateness of identifying CIT clusters utilizing a leukemia cohort from a large inpatient population treated with chemotherapy and hematopoietic stem cell transplants (HSCT).
We used US national inpatient data and applied appropriate weight to explore the original patient numbers. We applied the principal component analysis (PCA) with Multivariate Analysis with Optimal Scaling (MVAOS) method. MVAOS allows the categorical transformations computed concurrently with the PCA to maximize the variance explained by each component. PCA was employed to identify significant principal components from each treatment cohort, and we weighted each contributing variable. To produce CIT scores, variable-specific factor scores were applied to standardized Allogenic HSCT and Chemotherapy values. NIS are provided with numbers (unweighted) vs. Original numbers (weighted) we use weights to get original numbers in all our analysis. We are using further analysis classifying allogenic HSCT and chemotherapy cases/numbers and assigning numbers to it. (i.e., converting numerical values to logical and Nominal categories of the MVOAS).
1,658,134 leukemia patients, with a mean [SD] age 57.8 [20.4]. 69.9% were Whites, 12.9% Blacks, 11.2% Hispanics, and 5.9% Others. 71,780 toxicities with oral, gastric (GI), dermatologic, and constitutional symptoms were reported. MVOS-PCA (Fig 1) demonstrated positive correlation. Principle component analysis 1 of GI toxicities shows positive correlation, whereas PC2 (Fig 2) contrasts anorexia, nausea, constipation, and diarrhea (positive PC2) with distension-bloating, and dyspepsia as not being correlated. The proportion of transformed variance explained (Cumulative Variance Accounted For) by each PC, in this case (GI), was 18% for PC1 and 36% for PC2.
The application of MVOS-PCA to a large claims database successfully identified CIT clusters and confirmed the clinical observation that patients suffer multiple toxicities, often path biologically related, simultaneously. And this innovation enhance the adaptability of research and planning for supportive care, which is crucial for determining the optimal treatment approach for CIT.
Employing Explainable Artificial Intelligence (XAI) to Predict Cachexia Among Patients With Pancreatic or Oral Cavity Cancers. Hannah Wardill, Minu P. Mohan, Stephen Sonis, Joel Epstein, Sudheer B. K. Thazhe, Rhine Sukumar, Sreelakshmi Panginikkode, Venu Gopalakrishnan, Roberto Pili, Satheesh kumar Poolakkad Sankaran. Abstracts for MASCC/AFSOS/ISOO Annual Meeting 2024. Support Care Cancer 32 (Suppl 1), 434 (2024). https://doi.org/10.1007/s00520-024-08541-z
Future clinical decision-making is expected to be influenced by the application of machine learning (ML) and eXplainable Artificial Intelligence (XAI) techniques. XAI algorithms may be of particular value in informing the best supportive cancer care (SCC). Here, we utilized ML algorithms and XAI-based interpretably in predicting cachexia in patients with two solid cancers—pancreatic and oral cavity cancers (OC).
In this research, predictive analytics, XAI, and machine learning (ML) techniques were applied to the National Inpatient Sample (NIS) database. We employed linear and non-linear machine learning (ML) models, specifically Lasso, Ridge, and Random Forest (RF). Additionally, for XAI, we utilized Neural Networks (NN) and RF. The datasets were divided into training and test sets, with 70% allocated for training and 30% for testing. In addition, we employed XAI techniques such as permutation importance, global surrogate marker, feature interpretation, interaction, and local interpretability.
The cachexia study population consisted of 750 with pancreatic cancers and 1885 with OC. The feature selection methods successfully identified demographic, socioeconomic, and clinical variables. Linear models, Lasso and Ridge, showed an excellent area under the ROC for all cohorts (>0.9 in train and test datasets). In contrast, tree-based models, such as RF, only performed excellently in pancreatic cancer cohorts (AUC 0.99 for train and 0.92 for test), while in OC, this was 0.99 in train and 0.85 in test. The XAI techniques showed that those aged>70 years, Blacks, and those residing in low-income neighborhoods exhibited a higher likelihood of developing cachexia. The prevalence of weight loss was consistently coupled with drug addiction, congestive heart failure, and chronic pulmonary disease.
Coupled with the amount of data resulting from the digitalization of healthcare systems, the computational capabilities of modern servers to create ML-based care pathways pave the way for a revolution in SCC's clinical decision-making.
Stem cell transplant cancer patients experiencing oral mucositis is associated with infectious complications: systematic review and meta-analysis. Abstracts for MASCC/AFSOS/ISOO Annual Meeting 2024. Support Care Cancer 32 (Suppl 1), 434 (2024). https://doi.org/10.1007/s00520-024-08541-z
Authors: Susan Eichhorn, MD, MPH1; Roberto Pili, MD2; Joel B Epstein, DMD, MSD3; Shaniza Haniff, MD4; Sreelakshmi Panginikkode5, MD; Minu Ponnamma Mohan6 DDS, MDS, MPH; Poolakkad S Satheeshkumar, DDS, MDS, MSc, MMSc7*.
The oral cavity is recognized as a potential origin of systemic infections via the oral-systemic pathway. The existing body of evidence concerning the effects of oral mucositis (OM) induced by high-dose chemotherapy on the prognosis of systemic infections is extremely limited. Our objective is to quantify randomized, non-randomized, and observational studies that assess the systemic and local infectious complications associated with OM in cancer patients undergoing hematopoietic stem cell transplant (HSCT), and further assess risk factors for OM.
We conducted a comprehensive literature search using PubMed and EBSCO databases, from inception to June 13, 2021. Our search adhered to the Preferred Reporting Items for Systematic Review and Meta-Analyses (PRISMA) criteria. Information regarding risk factors for OM and the occurrence of systemic infectious complications related to OM was obtained from suitable studies utilizing a predetermined data extraction form. The assessment of risk of bias was conducted using the Newcastle-Ottawa Scale for non-randomized, and observational research, and the Cochrane Collaboration Tool for randomized control trials.
We identified 1,438 articles, after quality assessment, we synthesized 24 studies which were subjected to risk factor evaluation, and 4 (611 patients) studies were included in the meta-analysis of infectious complications outcomes. Among cancer patients who received HSCT, patients with OM have increased risk of developing infectious complications (OR Random=3.84, 05%CI, 2.51-5.86P, Heterogeneity: I2 = 0%, tau2 = 0, p = 0.96). Risk factors were categorized into baseline patient characteristics, laboratory results, cancer treatment and conditioning regimens, and OM prophylaxis. Identified OM risk factors included female sex, reduced kidney function, presence of HSV-1, longer duration of neutropenia, higher intensity conditioning, and methotrexate use.
The study's results suggest that preventing OM could be beneficial for cancer patients undergoing HSCT who are experiencing systemic infections. These findings indicate an association between OM and the development of infectious problems.