“Drug shortages can extend beyond leukemia-[specific] drugs,” Jayastu Senapati, MBBS, an assistant professor in the Department of Leukemia, Division of Cancer Medicine, at The University of Texas MD Anderson Cancer Center in Houston, said in an interview with OncLive®. “We have had shortages of normal saline and meperidine. These drugs are bread-and-butter in any internal medicine, leukemia, or oncology setup. I believe a lot of these need to be made closer to home so that these shortages can be averted.”
Oncologic drug shortages can be caused by a variety of factors, including low prices, quality assurance concerns, manufacturing complexities, and the concentration of production in only select areas.1 In hematologic oncology specifically, shortages of drugs that are used as parts of lymphodepleting regimens prior to CAR T-cell therapy, such as fludarabine can delay treatment for patients. Drug shortages can also lead to the use of less effective treatment options or missed doses of therapy.
How can hematologic oncologists adapt to drug shortages?
When faced with the specter of a drug shortage, it is important to understand which drugs can be suitably substituted without comprising patient safety or treatment efficacy, Harry P. Erba, MD, PhD, explained. “If you’ve already decided on what you believe is the best option for a patient, having a similar regimen is very important. In terms of patient safety, the most important issue when you're talking about acute leukemias is the rapid introduction of that treatment. Having the ability to switch to another drug is important,” Erba said in an interview with OncLive.
Erba is a professor of medicine in the Department of Hematologic Malignancies and Cellular Therapy at Duke Cancer Institute in Durham, North Carolina.
As of October 2025, azacitidine is listed on the FDA’s drug shortages tracker.2 Erba noted that although decitabine can be used as a substitute for azacitidine outside of clinical trials, the differences in the toxicity profiles of the agents must be taken into consideration.
Findings from a systematic review and network meta-analysis that included 8 clinical trials which enrolled patients withacute myeloid leukemia (AML) and higher-risk myelodysplastic syndromes (n = 2135) published in Frontiers in Pharmacology revealed that treatment with decitabine increased the risk of high-grade anemia (relative risk [RR], 1.61; 95% CI, 1.03-2.51), febrile neutropenia (RR, 4.03; 95% CI, 1.41-11.52), and leukopenia (RR, 3.43; 95% CI, 1.64-7.16) compared with azacitidine.3 “Sometimes, we need to adjust the dose of decitabine to make it less myelosuppressive,” Erba noted.
How could AI be used to mitigate drug shortages?
In the near future, artificial intelligence (AI)–enabled systems could represent a key tool in the battle against drug shortages in oncology, Senapati said. AI could potentially be used to increase drug shortage visibility, anticipate supply chain disruptions, and cut down on waste in the manufacturing process.4 These systems can account for factors such as market trends, weather, and geopolitical events to modify production schedules, identify alternate suppliers, and redistribute inventory in real time.
“Normal algorithms can predict what can happen over the next 3 to 6 months [but] there are so many other variables involved. AI could help in predicting, [for example], maybe 1.5 or 2 years later, that there will be an increase in patients [receiving] CAR T-cell therapy or that [since] we are not making as much fludarabine there could be a crunch. Predicting supply-demand mismatch will be an important area where AI could help,” Senapati explained.
For example, in a study published in the American Journal of Health-System Pharmacy, investigators developed and validated a predictive model to anticipate and manage drug shortages.5 The machine learning model was trained using k-fold cross-validation and a 70/30 partition into a training dataset and a testing dataset, respectively.
A total of 1517 drugs that were in shortage and not in shortage were included in the model dataset. Candidate predictive factors consisted of dosage form, therapeutic class, controlled substance schedule, orphan drug status, generic versus branded status, and number of manufacturers.
Results from the multiple logistic regression revealed that factors that positively predicted a drug shortage consisted of classification as an intravenous (IV)-only medication (OR, 3.94 95% CI, 1.43-4.13), both oral and IV medication (OR, 2.41; 95% CI, 1.43-4.13), antimicrobial (OR, 3.68; 95% CI, 2.28-5.98), analgesic (OR, 8.41; 95% CI, 4.00-18.1), electrolyte (OR, 101.1; 95% CI, 30.1-481.7), anesthetic (OR, 12.9; 95% CI, 3.17-65.3), or cardiovascular agent (OR, 1.90; 95% CI, 1.15-3.15). Findings from the validation of the model demonstrated that the sensitivity of the model was 0.71 and the specificity was 0.93. The overall accuracy of the model was 0.87.
“In places where community oncologists who might not have access to as many clinical trial options or [drug alternatives], AI could potentially help [to address drug shortages],” Senapati said. “AI cannot develop a regimen for a patient, but it could give options to the physician to see what other drugs could be used.”
References
- Malta M, Christian M, Cadwallader AB. Oncology drug shortages: impacts, policy reforms, and advocacy imperatives. Cancer J. 2025;31(5):e0784. doi:10.1097/PPO.0000000000000784
- FDA drug shortages. FDA. Accessed October 29, 2025. https://www.accessdata.fda.gov/scripts/drugshortages/dsp_ActiveIngredientDetails.cfm?AI=Azacitidine%20Injection&st=c&tab=tabs-1
- Ma J, Ge Z. Comparison between decitabine and azacitidine for patients with acute myeloid leukemia and higher-risk myelodysplastic syndrome: a systematic review and network meta-analysis. Front Pharmacol. 2021;12:701690. doi:10.3389/fphar.2021.701690
- Simpson MD, Qasim HS. Clinical and operational applications of artificial intelligence and machine learning in pharmacy: a narrative review of real-world applications. Pharmacy (Basel). 2025;13(2):41. doi:10.3390/pharmacy13020041
- Liu I, Colmenares E, Tak C, et al. Development and validation of a predictive model to predict and manage drug shortages. Am J Health Syst Pharm. 2021;78(14):1309-1316. doi:10.1093/ajhp/zxab152
