Date of Award

January 2024

Document Type

Open Access Thesis

Degree Name

Medical Doctor (MD)



First Advisor

Jonathan S. Leventhal

Second Advisor

Caroline A. Nelson


Dermatologic infections affecting the skin, hair, mucosae, and nails are common complications of systemic cancer therapy and can impact cancer patients’ quality of life; however, studies investigating these complications are largely lacking. We aimed to characterize dermatologic infectious complications secondary to oncologic therapy and the management of these complications in the outpatient and inpatient settings. We sought to identify how dermatologic infections differ in patients undergoing novel classes of cancer therapy, including targeted therapy and immunotherapy, compared to those on traditional cytotoxic chemotherapy. Second, we aimed to investigate non-infectious dermatologic complications of oncologic therapy which may mimic infectious complications. Here, we present the results of retrospective chart reviews of the Yale New Haven Health electronic medical record on dermatologic infectious complications as well as non-infectious mimickers in cancer patients on systemic cancer treatment. Joint Data Analytics Team queries as well as manual chart review were used to identify patients seen in the outpatient oncodermatology clinic and inpatient consultative service with a cancer diagnosis and a dermatologic infectious diagnosis or pseudocellulitis diagnosis. The majority of reported infections occurred in patients receiving regimens including cytotoxic therapy (67.4%). In comparison, fewer infections were reported in patients on targeted agents (27.2%) or immunotherapy (5.4%). We found that invasive bacterial infections, disseminated viral infections, invasive fungal infections, and atypical infections occurred most commonly in patients treated with regimens including cytotoxic therapy: of infections associated with cytotoxic therapy, 60.8% of bacterial infections and 6.8% of fungal infections were invasive, while 17.9% of viral infections were disseminated. Targeted therapy and immunotherapy were less often associated with severe soft tissue infections. Invasive or disseminated infections were also more likely to require inpatient management or impact cancer therapy. Therefore, there should be a low threshold for obtaining sterile tissue cultures for bacterial, fungal, and mycobacterial organisms in these patients to decrease the risk of life-threatening complications. Non-infectious dermatologic complications such as pseudocellulitis, on the other hand, may mimic infections. Distinguishing cellulitis from pseudocellulitis is important to minimize antibiotic usage and interruptions of cancer therapy for oncologic patients. We characterized dermatologic adverse effects of various cancer therapies that presented similarly to cellulitis and were often treated with antibiotics before consultation with dermatology. Acute lipodermatosclerosis was the most common condition presenting as pseudocellulitis (74.2%), associated with gemcitabine in 73.9% of cases. Other dermatologic diagnoses of pseudocellulitis included venous stasis dermatitis, injection site reaction/extravasation injury, and edema bullae, among others. 77.4% of patients received systemic antibiotics before referral to dermatology, emphasizing the difficulty of distinguishing it from true cellulitis and the decision to treat empirically with antibiotics in a high-risk population. 32.3% of patients experienced interruption of cancer therapy due to concern for cellulitis. After dermatologic consultation, pseudocellulitis was successfully treated with topical anti-inflammatory agents. Recognition of both infectious and non-infectious dermatologic complications of cancer therapy is essential for proper management and to minimize interruptions in life-preserving oncologic treatment.


This is an Open Access Thesis.

Open Access

This Article is Open Access