Differentiating Infection from Malignancy using SUVmax

Introduction to SUVmax Limitations

The American College of Chest Physicians recommends combining visual interpretation with quantitative values, rather than relying solely on SUVmax, to categorize PET findings as definitely benign, probably benign, indeterminate, probably malignant, or definitely malignant 1
SUVmax alone cannot reliably differentiate infection from malignancy, and relying on a single cutoff value is inadequate for clinical decision-making 1, 2

False-positive rates are extremely high in endemic regions, with PET-CT specificity dropping to as low as 25% due to granulomatous disease mimicking malignancy 1, 2
Visual assessment outperforms strict SUVmax cutoffs, with experienced readers achieving sensitivity of 96-100% and specificity of 76-86% using visual interpretation combined with clinical context 1, 2
Infectious lesions can show markedly elevated SUVmax, with tuberculosis, sarcoidosis, rheumatoid nodules, and acute inflammatory lesions demonstrating increased FDG uptake that overlaps with malignant values 1, 3

An SUVmax cutoff of 2.5 achieved 100% specificity and 86% sensitivity for detecting residual/recurrent lymphoma in one study of 27 patients 4, 5
One study found all SUVmax values >5.9 were malignant rather than infectious, but this was in a non-endemic TB region 2

Amino acid PET (FET, MET, FDOPA) is superior to FDG-PET for differentiating gliomas from non-neoplastic lesions, though moderately increased uptake can still occur in acute inflammatory lesions like active multiple sclerosis and brain abscesses 3
The American College of Chest Physicians recommends considering clinical context systematically, including geographic location, recent medication history, post-treatment timing, and morphologic CT features 1, 2, 4, 6

Do not use SUVmax as a standalone parameter in high-risk scenarios, such as TB-endemic regions, post-cytokine administration, post-therapy settings, and vertebral osteomyelitis 1, 2, 4, 5, 6
Pursue tissue diagnosis when uncertainty exists, particularly in TB-endemic regions, using integration of EBUS-TBNA or other biopsy methods 1, 2

PET‑CT frequently yields false‑positive results when foreign material is present, such as post‑implantation inflammatory reactions and infections related to indwelling devices. 7

In cases of prosthetic‑valve endocarditis—a scenario comparable to other implanted foreign bodies—SUVmax measurements of 7.1 and 7.8 have been documented as false‑positive findings caused by foreign‑body reaction and recent implantation. 7

Elevated inflammatory markers (e.g., C‑reactive protein, leukocyte count) enhance the sensitivity of PET imaging for detecting infections. 7