1
|
Lorentzen EH, Minami CA. Avoiding Locoregional Overtreatment in Older Adults With Early-Stage Breast Cancer. Clin Breast Cancer 2024; 24:319-327. [PMID: 38461117 PMCID: PMC11261391 DOI: 10.1016/j.clbc.2024.02.004] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/31/2023] [Revised: 02/01/2024] [Accepted: 02/07/2024] [Indexed: 03/11/2024]
Abstract
Advances in the treatment of older women with early-stage breast cancer, particularly opportunities for de-escalation of therapy, have afforded patients and providers opportunity to individualize care. As the majority of women ≥65 have estrogen receptor-positive, HER2-negative disease, locoregional therapy (surgery and/or radiation) may be tailored based on a patient's physiologic age to avoid either over- or undertreatment. To determine who would derive benefit from more or less intensive therapy, an accurate assessment of an older patient's physiologic age and incorporation of patient-specific values are paramount. While there now exist well-validated geriatric assessment tools whose use is encouraged by the American Society of Clinical Oncology when considering systemic therapy, these instruments have not been widely integrated into the locoregional breast cancer care model. This review aims to highlight the importance of assessing frailty and the concepts of and over- and undertreatment, in the context of trial data supporting opportunities for safe deescalation of locoregional therapy, when treating older women with early-stage breast cancer.
Collapse
Affiliation(s)
- Eliza H Lorentzen
- Department of Surgery, Brigham and Women's Hospital, Boston, MA; Center for Surgery and Public Health, Brigham and Women's Hospital, Boston, MA.
| | - Christina A Minami
- Department of Surgery, Brigham and Women's Hospital, Boston, MA; Center for Surgery and Public Health, Brigham and Women's Hospital, Boston, MA; Breast Oncology Program, Dana-Farber/Brigham and Women's Cancer Center, Boston, MA
| |
Collapse
|
2
|
Tasaka S, Kohada Y, Ikeda M, Kanaoka R, Hayashi M, Hinata N. Utility of the modified 5-item frailty index as a predictor of postoperative febrile urinary tract infection in patients who underwent ureteroscopy with laser lithotripsy. World J Urol 2024; 42:323. [PMID: 38748255 PMCID: PMC11096205 DOI: 10.1007/s00345-024-05016-y] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/12/2023] [Accepted: 04/22/2024] [Indexed: 05/18/2024] Open
Abstract
PURPOSE This study aimed to assess the effect of the modified 5-item frailty index on perioperative complications and surgical outcomes in patients who underwent ureteroscopy with laser lithotripsy for upper urinary tract stones. METHODS Patients who underwent ureteroscopy with laser lithotripsy for upper urinary tract stones between 2019 and 2022 were reviewed retrospectively. Assessment was performed using the modified 5-item frailty index based on medical history (hypertension, diabetes, heart failure, chronic obstructive pulmonary disease) and functional status. Patients were categorized into the high (≥ 2) and low (≤ 1) modified 5-item frailty index groups based on the frailty score. We compared the perioperative complications and surgical outcomes between the two groups. RESULTS Seventy-one (15.8%) and 393 (84.1%) of the 467 patients were classified into the high and low modified 5-item frailty index groups, respectively. The high modified 5-item frailty index group exhibited a significant association with increased febrile urinary tract infections compared to the low modified 5-item frailty index group [≥ 37.8 °C: 15 (20.3%) vs 13 (3.3%), p < 0.001; ≥ 38 °C: 9 (12.2%) vs 7 (1.8%), p < 0.001]. Surgical outcomes, including operative time and stone-free rate, did not differ significantly between the two groups. CONCLUSION The modified 5-item frailty index is valuable for predicting postoperative complications, particularly febrile urinary tract infections, after ureteroscopy with laser lithotripsy for upper urinary tract stones. This index allows for practical preoperative risk assessment in patients who underwent ureteroscopy with laser lithotripsy.
Collapse
Affiliation(s)
- Shinsaku Tasaka
- Department of Urology, Takanobashi Central Hospital, Hiroshima, Japan
- Department of Urology, Hiroshima University Graduate School of Biomedical Sciences, 1-2-3 Kasumi, Minamiku, Hiroshima, 734-8551, Japan
| | - Yuki Kohada
- Department of Urology, Hiroshima University Graduate School of Biomedical Sciences, 1-2-3 Kasumi, Minamiku, Hiroshima, 734-8551, Japan.
| | - Mikio Ikeda
- Department of Urology, Takanobashi Central Hospital, Hiroshima, Japan
| | - Ryuhei Kanaoka
- Department of Urology, Takanobashi Central Hospital, Hiroshima, Japan
| | - Mutsuo Hayashi
- Department of Urology, Takanobashi Central Hospital, Hiroshima, Japan
| | - Nobuyuki Hinata
- Department of Urology, Hiroshima University Graduate School of Biomedical Sciences, 1-2-3 Kasumi, Minamiku, Hiroshima, 734-8551, Japan
| |
Collapse
|
3
|
Agathis AZ, Bangla VG, Divino CM. Assessing the mFI-5 frailty score and functional status in geriatric patients undergoing inguinal hernia repairs. Hernia 2024; 28:135-145. [PMID: 37878113 DOI: 10.1007/s10029-023-02905-w] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/08/2023] [Accepted: 09/24/2023] [Indexed: 10/26/2023]
Abstract
PURPOSE The modified 5-factor frailty index (mFI-5) is a prognostic tool based on five comorbidities from the American College of Surgeons (ACS) National Surgical Quality Improvement Program (NSQIP) database-hypertension, congestive heart failure, chronic obstructive pulmonary disease (COPD), diabetes, and non-independent functional status. Our study investigates the mFI-5 index's ability to predict morbidity, length of stay (LOS), and discharge destination in geriatric patients undergoing inguinal hernia repairs, as well as assesses the interplay of baseline functional status. METHODS Patients aged ≥ 65 years who underwent inguinal or femoral hernia repairs from the 2018-2020 NSQIP database were studied. Separate analyses were performed for emergent and elective cohorts. Stratification was performed according to the sum of mFI-5 variables: mFI = 0, mFI = 1, mFI ≥ 2. RESULTS A total of 41,897 consisted of 92.9% elective and 7.1% emergent cases. The sample was 37.8% mFI = 0, 47.2% mFI = 1, and 15.0% mFI ≥ 2. Median age was 73 (IQR 68-78). Of emergent mFI ≥ 2 cases, 24.2% had non-independent functional status, versus only 4.8% in elective cases. Area under the curve was calculated for emergent and elective groups, including mortality (0.86, 0.80), pneumonia (0.82, 0.77), discharge destination not home (0.78, 0.73), prolonged LOS (0.69, 0.66), and infection (0.71, 0.62). Of index variables, dependent functional status was correlated with increased complications in elective and emergent cohorts, while COPD was significant in elective cases (OR > 2.0, p < 0.05). CONCLUSION The mFI-5 is predictive of complications in geriatric inguinal hernia repairs, especially in emergent cases. Frail patients with non-independent functional status are most at risk and, thus require proactive and watchful perioperative care.
Collapse
Affiliation(s)
- A Z Agathis
- Division of General Surgery, Department of Surgery, Icahn School of Medicine at Mount Sinai, One Gustave L. Levy Place, Box 1259, New York, NY, 10029, USA
| | - V G Bangla
- Division of General Surgery, Department of Surgery, Icahn School of Medicine at Mount Sinai, One Gustave L. Levy Place, Box 1259, New York, NY, 10029, USA
| | - C M Divino
- Division of General Surgery, Department of Surgery, Icahn School of Medicine at Mount Sinai, One Gustave L. Levy Place, Box 1259, New York, NY, 10029, USA.
| |
Collapse
|
4
|
Tasman J, Roberson PNE, Clegg D, Boukovalas S, Lloyd J. The impact of rural structural and community health factors on postmastectomy complications among south central Appalachian breast cancer patients. J Rural Health 2024; 40:104-113. [PMID: 37144973 DOI: 10.1111/jrh.12766] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/28/2022] [Revised: 04/17/2023] [Accepted: 04/27/2023] [Indexed: 05/06/2023]
Abstract
PURPOSE The study examined how structural and community health factors, including primary care physicians (PCP), food insecurity, diabetes, and mortality rate per county, are linked to the number and severity of postmastectomy complications among south central Appalachian breast cancer patients depending on rural status. METHODS Data was obtained through a retrospective review of 473 breast cancer patients that underwent a mastectomy from 2017 to 2021. Patient's ZIP Code was used to determine their rural-urban community area code and their county of residence for census data. We conducted a zero inflated Poisson regression. FINDINGS Results demonstrated that patients in small rural/isolated areas with low (B = -4.10, SE = 1.93, OR = 0.02, p = 0.03) to average (B = -2.67, SE = 1.32, OR = 0.07, p = 0.04) food insecurity and average (B = -2.67, SE = 1.32, OR = 0.07, p = 0.04) to high (B = -10.62, SE = 4.71, OR = 0.00, p = 0.02) PCP have significantly fewer postmastectomy complications compared to their urban counterparts. Additionally, patients residing in small rural/isolated areas with high (B = 4.47, SE = 0.49, d = 0.42, p < 0.001) diabetes and low mortality (B = 5.70, SE = 0.58, d = 0.45, p < 0.001) rates have significantly more severe postmastectomy complications. CONCLUSION These findings demonstrate that patients who reside in small/rural isolated areas may experience fewer and less severe postmastectomy when there is certain optimal structural and community health factors present compared to their urban counterparts. Oncologic care teams could utilize this information in routine consult for risk assessment and mitigation. Future research should further examine additional risks for postmastectomy complications.
Collapse
Affiliation(s)
- Jordan Tasman
- College of Nursing, The University of Tennessee, Knoxville, Tennessee, USA
| | | | - Devin Clegg
- Department of Surgery, University of Tennessee Graduate School of Medicine, Knoxville, Tennessee, USA
| | - Stefanos Boukovalas
- Division of Plastic Surgery, The University of Tennessee Graduate School of Medicine, Knoxville, Tennessee, USA
| | - Jillian Lloyd
- Cancer Institute, The University of Tennessee Graduate School of Medicine, Knoxville, Tennessee, USA
| |
Collapse
|
5
|
Agathis AZ, Bangla VG, Divino CM. Role of mFI-5 in predicting geriatric outcomes in laparoscopic cholecystectomy. Am J Surg 2023; 226:697-702. [PMID: 37633764 DOI: 10.1016/j.amjsurg.2023.07.039] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/26/2023] [Revised: 07/23/2023] [Accepted: 07/27/2023] [Indexed: 08/28/2023]
Abstract
BACKGROUND Frailty is the age-related decline contributing to adverse outcome vulnerability. This study assesses the modified 5-factor frailty index's (mFI-5) ability to predict geriatric cholecystectomy outcomes. METHODS Laparoscopic cholecystectomy patients ages ≥65 were identified from the American College of Surgeons' National Surgical Quality Improvement Program database (2018-2020). MFI-5 variables include hypertension, congestive heart failure, chronic obstructive pulmonary disease, diabetes, and functional status. Groups were stratified according to the number of comorbidities: mFI = 0, mFI = 1, mFI≥2. RESULTS 32,481 cases included 27.6% mFI = 0, 46.4% mFI = 1, 26.0% mFI≥2. Highest frailty correlated with increased discharges to not home (OR 1.88, p < 0.01). Non-independent functional status was associated with mortality (OR 7.32), prolonged length of stay (LOS) (5.69), pneumonia (4.90), sepsis (3.78), readmission (2.60) (p < 0.01). AUCs were calculated for prolonged LOS (0.89), discharges to not home (0.85), mortality (0.83), pneumonia (0.76), sepsis (0.76). CONCLUSIONS Healthcare teams can use mFI-5 to target at-risk cholecystectomy patients and proactively intervene to avoid complications.
Collapse
Affiliation(s)
- Alexandra Z Agathis
- Division of General Surgery, Department of Surgery, Icahn School of Medicine at Mount Sinai, New York, NY, USA
| | - Venu G Bangla
- Division of General Surgery, Department of Surgery, Icahn School of Medicine at Mount Sinai, New York, NY, USA
| | - Celia M Divino
- Division of General Surgery, Department of Surgery, Icahn School of Medicine at Mount Sinai, New York, NY, USA.
| |
Collapse
|
6
|
Clements NA, Gaskins JT, Martin RCG. Predictive Ability of Comorbidity Indices for Surgical Morbidity and Mortality: a Systematic Review and Meta-analysis. J Gastrointest Surg 2023; 27:1971-1987. [PMID: 37430092 DOI: 10.1007/s11605-023-05743-4] [Citation(s) in RCA: 1] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 03/07/2023] [Accepted: 05/27/2023] [Indexed: 07/12/2023]
Abstract
BACKGROUND Several contemporary risk stratification tools are now being used since the development of the Charlson Comorbidity Index (CCI) in 1987. The purpose of this systematic review and meta-analysis was to compare the utility of commonly used co-morbidity indices in predicting surgical outcomes. METHODS A comprehensive review was performed to identify studies reporting an association between a pre-operative co-morbidity measurement and an outcome (30-day/in-hospital morbidity/mortality, 90-day morbidity/mortality, and severe complications). Meta-analysis was performed on the pooled data. RESULTS A total of 111 included studies were included with a total cohort size 25,011,834 patients. The studies reporting the 5-item Modified Frailty Index (mFI-5) demonstrated a statistical association with an increase in the odds of in-hospital/30-day mortality (OR:1.97,95%CI: 1.55-2.49, p < 0.01). The pooled CCI results demonstrated an increase in the odds for in-hospital/30-day mortality (OR:1.44,95%CI: 1.27-1.64, p < 0.01). Pooled results for co-morbidity indices utilizing a scale-based continuous predictor were significantly associated with an increase in the odds of in-hospital/30-day morbidity (OR:1.32, 95% CI: 1.20-1.46, p < 0.01). On pooled analysis, the categorical results showed a higher odd for in-hospital/30-day morbidity (OR:1.74,95% CI: 1.50-2.02, p < 0.01). The mFI-5 was significantly associated with severe complications (Clavien-Dindo ≥ III) (OR:3.31,95% CI:1.13-9.67, p < 0.04). Pooled results for CCI showed a positive trend toward severe complications but were not significant. CONCLUSION The contemporary frailty-based index, mFI-5, outperformed the CCI in predicting short-term mortality and severe complications post-surgically. Risk stratification instruments that include a measure of frailty may be more predictive of surgical outcomes compared to traditional indices like the CCI.
Collapse
Affiliation(s)
- Noah A Clements
- The Hiram C. Polk, Jr., MD Department of Surgery, Division of Surgical Oncology, University of Louisville School of Medicine, 315 E. Broadway, Louisville, KY, 40292, USA
| | - Jeremy T Gaskins
- The Hiram C. Polk, Jr., MD Department of Surgery, Division of Surgical Oncology, University of Louisville School of Medicine, 315 E. Broadway, Louisville, KY, 40292, USA
| | - Robert C G Martin
- The Hiram C. Polk, Jr., MD Department of Surgery, Division of Surgical Oncology, University of Louisville School of Medicine, 315 E. Broadway, Louisville, KY, 40292, USA.
| |
Collapse
|
7
|
Liu H, Akhavan A, Yin R, Ibelli T, Mandelbaum M, Katz A, Etigunta S, Alerte E, Kuruvilla A, Liu C, Taub PJ. Efficacy of the Modified 5-Item Frailty Index in Predicting Surgical-Site Infections in Patients Undergoing Breast Implant Augmentation: A National Surgical Quality Improvement Project-Based 5-Year Study. Aesthet Surg J Open Forum 2023; 5:ojad067. [PMID: 37575888 PMCID: PMC10413997 DOI: 10.1093/asjof/ojad067] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 08/15/2023] Open
Abstract
Background The ability to predict breast implant augmentation complications can significantly inform patient management. A frailty measure, such as the modified 5-item frailty index (mFI-5), is becoming an increasingly established risk factor for adverse postoperative outcomes. The authors hypothesized that the mFI-5 is predictive of 30-day postoperative complications in breast augmentation. Objectives To investigate if mFI-5 can predict the likelihood and magnitude of 30-day complications resulting from breast augmentations. Methods A retrospective review study of the National Surgical Quality Improvement Program database for patients who underwent breast implant augmentation without other concurrent procedures, from 2015 to 2019. Age, BMI, number of major comorbidities, American Society of Anesthesiologists (ASA) classifications, smoking status, mFI-5 score, and modified Charlson comorbidity index score were compared as predictors of all-cause 30-day complications and 30-day surgical-site complications using regression analyses. Results Overall, 2478 patients were analyzed, and among them, 53 patients developed complications (2.14%). mFI-5 score significantly predicted surgical-site infection (SSI) complications (odds ratio [OR] = 4.24, P = .026). Frail patients had a higher occurrence of SSIs than nonfrail patients (P = .049). Multivariable analyses showed ASA class predicted 30-day SSI complications (OR = 5.77, P = .027) and mFI-5 approached, but did not reach full significance in predicting overall 30-day complications (OR = 3.14, P = .085). Conclusions To date, the impact of frailty on breast implant procedure outcomes has not been studied. Our analysis demonstrates that the mFI-5 is a significant predictor for SSIs in breast implant augmentation surgery and is associated with overall complications. By preoperatively identifying frail patients, the surgical team can better account for postoperative support to minimize the risk of complications. Level of Evidence 4
Collapse
Affiliation(s)
| | | | | | | | | | | | | | | | | | - Chuanju Liu
- Corresponding Author: Dr Chuanju Liu, P.O. Box 208071, New Haven, CT 06520-8071, USA. E-mail:
| | | |
Collapse
|
8
|
Obed D, Knoedler S, Salim M, Gulbis N, Dastagir N, Dastagir K, Bingöl AS, Vogt PM. The modified 5-item frailty index as a predictor of complications in burn patients. JPRAS Open 2023; 36:62-71. [PMID: 37179743 PMCID: PMC10172613 DOI: 10.1016/j.jpra.2023.03.001] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 02/17/2023] [Accepted: 03/05/2023] [Indexed: 03/16/2023] Open
Abstract
The modified 5-item frailty index (mFI-5), as a measure of frailty and biological age, has been shown to be a reliable predictor of complications and mortality in a variety of surgical specialties. However, its role in burn care remains to be fully elucidated. We, therefore, correlated frailty with in-hospital mortality and complications after burn injury. The medical charts of all burn patients admitted between 2007 and 2020 who had ≥ 10 % of their total body surface area affected were retrospectively reviewed. Data on clinical, demographic, and outcome parameters were collected and evaluated, and mFI-5 was calculated on the basis of the data obtained. Univariate and multivariate regression analyses were used to investigate the association between mFI-5 and medical complications and in-hospital mortality. A total of 617 burn patients were included in this study. Increasing mFI-5 scores were significantly associated with increased in-hospital mortality (p < 0.0001), myocardial infarction (p = 0.03), sepsis (p = 0.005), urinary tract infections (p = 0.006), and perioperative blood transfusions (p = 0.0004). They were also associated with an increase in the length of hospital stay and the number of surgical procedures, albeit without statistical significance. An mFI-5 score of ≥ 2 was a significant predictor of sepsis (odds ratio [OR] = 2.08; 95% confidence interval [CI]: 1.03 to 3.95; p = 0.04), urinary tract infection (OR = 2.82; 95% CI: 1.47 to 5.19; p = 0.002), and perioperative blood transfusions (OR = 2.61; 95% CI: 1.61 to 4.25; p = 0.0001). Multivariate logistic regression analysis revealed that an mFI-5 score of ≥ 2 was not an independent risk factor for in-hospital mortality (OR = 1.44; 95% CI: 0.61 to 3.37; p = 0.40). mFI-5 is a significant risk factor for only a few select complications in the burn population. It is not a reliable predictor of in-hospital mortality. Therefore, its utility as a risk stratification tool in the burn unit may be limited.
Collapse
|
9
|
Chow AL, Karius AK, Broderick KP, Cooney CM. Frailty is the New Age: A Retrospective Study of Modified Frailty Index for Preoperative Risk Assessment in Autologous Breast Reconstruction. J Reconstr Microsurg 2023; 39:81-91. [PMID: 36691382 DOI: 10.1055/s-0042-1743566] [Citation(s) in RCA: 1] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 01/25/2023]
Abstract
BACKGROUND Age is a poor predictor of postoperative outcomes in breast reconstruction necessitating new methods for risk-stratifying patients preoperatively. The 5-item modified frailty index (mFI-5) is a validated measure of frailty which assesses patients' global health. The purpose of this study was to compare the effectiveness of mFI-5 and age as independent predictors of 30-day postoperative complications following autologous breast reconstruction. METHODS Patients who underwent autologous breast reconstruction between 2005 and 2019 were identified from the American College of Surgeons National Surgical Quality Improvement Program. Patients were stratified based on presence of major, minor, both minor and major, and no complications. Univariate and multivariate logistic regression were performed to determine the predictive power of mFI-5, age, and other preoperative risk factors for development of minor and major 30-day postoperative complications in all patients and stratified by flap type. RESULTS A total of 25,215 patients were included: 20,366 (80.8%) had no complications, 2,009 (8.0%) had minor complications, 1,531 (6.1%) had major complications, and 1,309 (5.2%) had both minor and major complications. Multivariate regression demonstrated age was not a predictor of minor (odds ratio [OR]: 1.0, p = 0.045), major (OR: 1.0, p = 0.367), or both minor and major (OR: 1.0, p = 0.908) postoperative complications. mFI-5 was a significant predictor of minor complications for mFI-5 scores 1 (OR: 1.3, p < 0.001), 2 (OR: 1.8, p < 0.001), and 3 (OR: 2.8, p = 0.043). For major complications, mFI-5 was a significant predictor for scores 1 (OR: 1.2, p = 0.011) and 2 (OR: 1.3, p = 0.03). CONCLUSION Compared with age, mFI-5 scores were better predictors of 30-day postoperative complications following autologous breast reconstruction regardless of flap type. Additionally, higher mFI-5 scores were associated with increased odds of minor and major complications. Our findings indicate that reconstructive breast surgeons should consider using the mFI-5 in lieu of age to risk-stratify patients prior to autologous breast reconstruction surgery.
Collapse
Affiliation(s)
- Amanda L Chow
- Department of Plastic and Reconstructive Surgery, The Johns Hopkins University School of Medicine, Baltimore, Maryland.,Division of Plastic and Reconstructive Surgery, Rutgers New Jersey Medical School, Newark, New Jersey
| | - Alexander K Karius
- Department of Plastic and Reconstructive Surgery, The Johns Hopkins University School of Medicine, Baltimore, Maryland
| | - Kristen P Broderick
- Department of Plastic and Reconstructive Surgery, The Johns Hopkins University School of Medicine, Baltimore, Maryland
| | - Carisa M Cooney
- Department of Plastic and Reconstructive Surgery, The Johns Hopkins University School of Medicine, Baltimore, Maryland
| |
Collapse
|
10
|
Qiao EM, Qian AS, Nalawade V, Voora RS, Kotha NV, Vitzthum LK, Murphy JD. Evaluating High-Dimensional Machine Learning Models to Predict Hospital Mortality Among Older Patients With Cancer. JCO Clin Cancer Inform 2022; 6:e2100186. [PMID: 35671416 DOI: 10.1200/cci.21.00186] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/20/2022] Open
Abstract
PURPOSE Older hospitalized cancer patients face high risks of hospital mortality. Improved risk stratification could help identify high-risk patients who may benefit from future interventions, although we lack validated tools to predict in-hospital mortality for patients with cancer. We evaluated the ability of a high-dimensional machine learning prediction model to predict inpatient mortality and compared the performance of this model to existing prediction indices. METHODS We identified patients with cancer older than 75 years from the National Emergency Department Sample between 2016 and 2018. We constructed a high-dimensional predictive model called Cancer Frailty Assessment Tool (cFAST), which used an extreme gradient boosting algorithm to predict in-hospital mortality. cFAST model inputs included patient demographic, hospital variables, and diagnosis codes. Model performance was assessed with an area under the curve (AUC) from receiver operating characteristic curves, with an AUC of 1.0 indicating perfect prediction. We compared model performance to existing indices including the Modified 5-Item Frailty Index, Charlson comorbidity index, and Hospital Frailty Risk Score. RESULTS We identified 2,723,330 weighted emergency department visits among older patients with cancer, of whom 144,653 (5.3%) died in the hospital. Our cFAST model included 240 features and demonstrated an AUC of 0.92. Comparator models including the Modified 5-Item Frailty Index, Charlson comorbidity index, and Hospital Frailty Risk Score achieved AUCs of 0.58, 0.62, and 0.71, respectively. Predictive features of the cFAST model included acute conditions (respiratory failure and shock), chronic conditions (lipidemia and hypertension), patient demographics (age and sex), and cancer and treatment characteristics (metastasis and palliative care). CONCLUSION High-dimensional machine learning models enabled accurate prediction of in-hospital mortality among older patients with cancer, outperforming existing prediction indices. These models show promise in identifying patients at risk of severe adverse outcomes, although additional validation and research studying clinical implementation of these tools is needed.
Collapse
Affiliation(s)
- Edmund M Qiao
- Department of Radiation Medicine and Applied Sciences, University of California, San Diego, La Jolla, CA
| | - Alexander S Qian
- Department of Radiation Medicine and Applied Sciences, University of California, San Diego, La Jolla, CA
| | - Vinit Nalawade
- Department of Radiation Medicine and Applied Sciences, University of California, San Diego, La Jolla, CA
| | - Rohith S Voora
- Department of Radiation Medicine and Applied Sciences, University of California, San Diego, La Jolla, CA
| | - Nikhil V Kotha
- Department of Radiation Medicine and Applied Sciences, University of California, San Diego, La Jolla, CA
| | - Lucas K Vitzthum
- Department of Radiation Oncology, Stanford University, Stanford, CA
| | - James D Murphy
- Department of Radiation Medicine and Applied Sciences, University of California, San Diego, La Jolla, CA
| |
Collapse
|
11
|
Patel SA, Baltodano PA. Editorial on predicting postoperative complications following mastectomy in the elderly: Evidence for the 5-factor frailty index. Breast J 2021; 27:507-508. [PMID: 34053164 DOI: 10.1111/tbj.14257] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/28/2022]
Affiliation(s)
- Sameer A Patel
- Division of Plastic and Reconstructive Surgery, Department of Surgical Oncology, Fox Chase Cancer Center, Philadelphia, Pennsylvania, USA
| | - Pablo A Baltodano
- Division of Plastic and Reconstructive Surgery, Department of Surgical Oncology, Fox Chase Cancer Center, Philadelphia, Pennsylvania, USA
| |
Collapse
|