Dear Editor,

We read with great interest the pictorial essay titled “Drug-induced lung disease: a brief update for radiologists” by Ufuk et al.¹ in the January 2023 of Diagnostic and Interventional Radiology issue. In this article, the authors comprehensively covered the imaging patterns and differential diagnoses of drug-induced lung disease (DILD) with demonstrative case examples.¹ We would like to mention a few points that may contribute to this article, which focuses on this complex issue, a popular topic in radiology.

First, according to the literature, a pulmonary hemorrhage can be classified as a radiological finding of DILD. A pulmonary hemorrhage can be seen as ground-glass opacities (GGO) or consolidations and is known to be associated with drugs such as penicillamine, rituximab, and cocaine.²

Second, Sridhar et al.³ proposed an alternative imaging-based classification system for DILDs consisting of six computed tomography patterns. Although this classification system overlaps with the patterns stated in the article, some differences draw attention.¹ “Organizing pneumonia” and “sarcoid-like” patterns show significant overlap. However, “fibrotic”, “diffuse GGO”, and “centrilobular ground-glass nodule (GGN)” patterns are suggested instead of “non-specific interstitial pneumonia”, “diffuse alveolar damage”, and “hypersensitivity pneumonitis” patterns, respectively.^1,3 Last, the “linear-septal” pattern indicates the findings of interstitial pulmonary edema.³

As an oncological lung imaging center, we have seen the centrilobular GGN pattern in several acute myeloid leukemia (AML) cases receiving intermediate/high dose cytosine arabinoside (Ara-C) chemotherapy, and it causes a differential diagnosis problem. Therefore, we want to share our experience on this matter through two cases. Although Ara-C pulmonary toxicity mainly presents with non-cardiogenic pulmonary edema, in 2009, Chagnon et al.⁴ described a new pattern characterized by bilateral diffuse/upper lobe-predominant centrilobular nodules in six febrile neutropenic AML patients who had recently received Ara-C (Figure 1).

We believe that awareness of this pattern can assist radiologists in shaping patient management. However, opportunistic infections, especially miliary tuberculosis, should always be considered in febrile neutropenia. Although miliary tuberculosis is generally characterized by randomly distributed uniform micronodules with sharper contours, relying solely on radiological findings might have devastating consequences. Therefore, as emphasized by Ufuk et al.¹, multidisciplinary meetings are vital, especially to establish a diagnosis and determine the best management plan for patients with suspected DILD. Nevertheless, even with a multidisciplinary approach, it is not always possible to make a definitive diagnosis due to various confounders, including radiological mimickers, empirical treatments, limitations of microbiological/serological examinations, comorbidities, and variability in the drug-effect temporal relationship (Figure 2). Thus, it is important to keep in mind that proper patient management is far more crucial than the definitive diagnosis.

Conflict of interest disclosure

The authors declared no conflicts of interest.