Cavitary lung lesions caused by Pneumocystis jirovecii in a patient with myelofibrosis on ruxolitinib

Summary

We report a rare case of a patient with Janus kinase 2-positive myelofibrosis on ruxolitinib, presenting with indolent pneumonia and cavitary lung lesions. Initial transthoracic biopsy was non-specific, but thoracoscopic biopsy revealed necrotising granulomatous disease caused by Pneumocystis jirovecii pneumonia (PJP). The patient, initially treated with trimethoprim–sulfamethoxazole, was switched to atovaquone due to gastrointestinal intolerance. Given the patient’s immunosuppression and extensive cavitary lesions, an extended course of atovaquone was administered, guided by serial imaging, resulting in clinical and radiological improvement. Unfortunately, the patient later passed away from a severe SARS-CoV-2 infection before complete radiographic resolution was observed. This case highlights the importance of recognising atypical PJP presentations causing granulomatous disease in immunosuppressed patients. While rare, documenting such cases may improve diagnosis using less invasive methods and help determine optimal treatment durations for resolution of these atypical infections.

Background

Pneumocystis jirovecii pneumonia (PJP) is a life-threatening pulmonary infection that predominantly affects immunocompromised individuals. In patients with HIV, the risk factors and typical presenting signs of PJP are relatively well understood, facilitating improved prevention and treatment strategies. However, in non-HIV-infected individuals, the risk, presentation and management of PJP are less known, leading to potential delays in diagnosis and treatment.^{1 2}

Patients with myelofibrosis treated with Janus kinase inhibitors constitute one such immunocompromised group in which the presentation and optimal management of PJP remain uncertain. Myelofibrosis contributes to an immunosuppressive state due to cell line disturbances, a reduced capacity for infection recovery and diminished spleen function.³ Furthermore, Janus kinase inhibitors such as ruxolitinib used to treat myelofibrosis may independently affect infection risk by downregulating critical immune and inflammatory signalling pathways and disrupting the activation of numerous immune cells.⁴ Ruxolitinib has been associated with multiple opportunistic infections, including PJP, cryptococcus, aspergillosis, non-tuberculous mycobacteria and cytomegalovirus.^{5 6}

Although PJP typically presents more acutely in non-HIV-infected individuals, indolent presentations are now recognised in patients with only mild-to-moderate immunosuppression. Specifically, PJP may present with nodular or cavitary lesions caused by granulomatous inflammation in such patients. Enhancing the recognition of atypical PJP presentations in patients with various degrees of immunosuppression is crucial due to the increasing prevalence of immunosuppression and resultant opportunistic infections.⁷,⁹

Case presentation

A female patient in her mid-60s with well-controlled type 2 diabetes mellitus and JAK-2-positive postpolycythemic myelofibrosis presented with a persistent cough, night sweats and fatigue progressively worsening over the prior 2 months. On examination, she appeared fatigued and mildly dyspneic with a temperature of 38.0°C. Her vital signs were otherwise normal with blood pressure 133/84 mm Hg, pulse 88 beats per minute, respiratory rate 18 breaths per minute and oxygen saturation 98% on room air. Lung auscultation revealed clear breath sounds bilaterally with no wheezes, rales or rhonchi. Her laboratory workup showed a white blood cell count of 23.1×10⁹/L (85% myeloid), haemoglobin of 111 g/L (normal 120–160), platelets of 220×10⁹/L (normal 150–450×10⁹/L), lactate dehydrogenase of 345 international units/L (normal 140–271), IgG of 281 mg/dL (normal 635–1741), IgA of 25 mg/dL (normal 66–433) and IgM was<20 mg/dL (normal 45–281). She was treated with ruxolitinib (20 mg two times per day) for 9 years. Severe cytopenias complicated her treatment course, along with recurrent left lower extremity ulcers due to varicella zoster, abundant recurrent human papillomavirus warts on her hands, recurrent vaginal intraepithelial neoplasia and an episode of disseminated herpes simplex virus.

Investigations

An X-ray of the chest revealed patchy bilateral infiltrates with multiple new lung masses (figure 1), leading to her admission for pneumonia. High-resolution CT of the chest showed numerous masses throughout both lung fields, with the largest in the upper lobes with evidence of cavitation (figure 2A). Further initial workup included blood cultures, which were negative after 5 days of incubation. Urine antigen tests for Streptococcus pneumoniae and Legionella pneumoniae were negative, as was methicillin-resistant Staphylococcus nasal screening. Tuberculosis was initially ruled out with acid-fast bacilli (AFB) smear and culture, all of which were negative on three separate sputum samples collected during the initial hospitalisation. She was discharged home with a short course of antibiotics for community-acquired pneumonia, which could not be completely ruled out despite the atypical presentation and imaging findings.

Figure 1. X-ray of the chest showing new patchy bilateral infiltrates and multiple lung masses.

Figure 2. Axial section from CT scans of the chest showing (A) bilateral infiltrates and multiple cavitary lesions at initial diagnosis and (B) after 4 months of treatment with atovaquone.

The patient remained symptomatic, and a follow-up percutaneous right upper lobe biopsy performed 15 weeks after initial presentation showed rare loosely necrotising granulomas. However, a detailed pathologic workup was otherwise non-specific. The biopsy results were reviewed by pathologists at a tertiary care centre in a specialised thoracic pathology consensus conference, which confirmed the non-diagnostic nature of the findings. A repeat transthoracic or transbronchial biopsy was considered. However, given the diffuse nature of the lung lesions and the wish for a prompt diagnosis after persistent debilitating symptoms, a shared decision was made to pursue video-assisted thoracoscopic surgery. This surgery was performed 27 weeks after the initial presentation and included a wedge resection of a portion of the lingula.

Perioperative laboratory workup prior to the thoracoscopy showed negative galactomannan and β-D-glucan was >500 pg/mL (normal: <60). The antinuclear antibody titre was <1:80, and immunofluorescence studies were negative for cytoplasmic or perinuclear antineutrophilic autoantibodies. The patient experienced a transient episode of hypercalcaemia for 2 weeks following the wedge resection procedure, with total calcium reaching a maximum of 11.9 mg/dL, parathyroid hormone (PTH) at 7 pg/mL (normal: 12–88), 25-hydroxy vitamin D at 46 ng/mL (normal: 30–95), and ionised calcium at 1.4 mmol/mL (normal: 1.2–1.38).

The wedge resection pathology revealed acute organising bronchopneumonia with necrotising granulomatous inflammation and foamy eosinophilic proteinaceous material (figure 3A). Periodic acid-Schiff, mucicarmine and AFB stains were negative. Grocott’s methenamine stain showed oval, crescent and helmet-shaped non-budding yeast forms (4–5 µm), most suggestive of PJP (figure 3B). Mycobacterium culture of lung tissue was negative as were three repeat AFB smears and cultures from sputum. Additionally, 1 week following discharge after the wedge resection, an outpatient induced sputum sample was obtained that was negative for Mycobacterium tuberculosis via nucleic acid amplification testing.

Figure 3. H&E stain showing granulomatous lung tissue from lingular specimen from patient (A) and Grocott’s methenamine stain from same tissue showing non-budding yeast forms consistent with Pneumocystis jirovecii pneumonia (B).

Differential diagnosis

The differential diagnosis for cavitary lung lesions is broad, encompassing malignancies, autoimmune causes and infectious diseases.¹⁰ The presence of multiple nodules and the patient’s immunocompromised status initially raised concerns for metastatic disease and lymphoma. However, the granulomatous tissue results from the initial lung biopsy made neoplasms, such as lymphoma less likely. Vasculitides, such as granulomatous polyangiitis, were considered less likely based on negative antinuclear antibody and antineutrophil cytoplasmic autoantibody tests. Sarcoidosis was also deemed less likely due to the absence of significant adenopathy in the chest. Fungal infections were highly suspected due to the patient’s high β-D-glucan level. Mycobacterial infections were also strongly considered due to their association with ruxolitinib.⁴ A thorough pathologic examination assisted in narrowing our diagnosis, with mycobacterium assessed to be unlikely based on multiple negative mycobacterium cultures; histoplasmosis was less likely due to the absence of intracellular organisms; blastomycosis and coccidioidomycosis were less likely given smaller size of organisms present; Candida and Aspergillus were less likely as no budding or hyphae were seen; and Cryptococcus was less likely based on negative periodic acid-Schiff and mucicarmine staining results.

Treatment

Immediately after diagnosis of PJP on biopsy, the patient started on trimethoprim–sulfamethoxazole (TMP-SMX) 160/800 mg three times per day. Concurrently, ruxolitinib was tapered to 5 mg two times per day due to cytopenias. Two intravenous immunoglobulin infusions were attempted but discontinued due to severe chills during administration. Given the patient’s immunocompromised state and the lack of guidelines for treating cavitary PJP lesions, a prolonged course of antibiotics was planned with interval imaging to monitor granuloma resolution.

Initial treatment with TMP-SMX proved unsuccessful due to severe nausea and vomiting, significantly limiting the patient’s ability to tolerate the medication. A repeat CT scan of the chest after 4 weeks of attempted treatment with TMP-SMX showed no change in the number or size of cavitary lesions. Consequently, treatment was switched to atovaquone 750 mg twice daily. The patient tolerated atovaquone well without significant gastrointestinal side effects and her symptoms began to improve significantly within days to short weeks after starting atovaquone therapy.

Outcome and follow-up

After 9 weeks of atovaquone treatment, imaging revealed resolution of cavitation and reduced pulmonary nodules (figure 2B). Despite this improvement, significant lung involvement persisted, prompting an extension of treatment for an additional 8 weeks. The patient continued to improve symptomatically and never required supplemental oxygen. Unfortunately, 26 weeks after initiating atovaquone treatment, just as the patient was becoming more active in the community, she contracted SARS-CoV-2. This additional infection led to severe respiratory failure, necessitating intensive care and ventilatory support. CT angiography of the chest showed extensive bilateral multifocal ground glass and consolidative airspace diseases; no pulmonary nodules were observed on the CT scan, though extensive ground glass opacities may have obscured them. Bronchoalveolar lavage revealed no Pneumocystis or fungal elements on periodic acid-Schiff or grocott’s methenamine stains and a negative Pneumocystis PCR. Despite aggressive treatment for COVID-19, the patient developed multiorgan failure and passed away after transitioning to comfort-focused care.

Discussion

Review of similar cases

PJP typically presents with symmetric, diffuse interstitial infiltrates but, on rare occasions, can form nodular or cavitary lesions,¹⁰ including necrotising granulomas.¹¹,¹⁷ We have identified two previous case reports of PJP associated with ruxolitinib,^{18 19} both of which report more common findings of interstitial infiltrates. Although Lee et al¹⁹ also reported nodular changes in a patient with PJP on ruxolitinib, to our knowledge, this is the first case report of confirmed necrotising granulomatous disease associated with myelofibrosis or ruxolitinib.

Pathophysiology

Granuloma formation from PJP is suspected to occur first as nodules that may progress to cavitary lesions through ischaemic necrosis from the invasion of Pneumocystis into vascular lumina or elastolytic proteases liberated by activated macrophages. The formation is generally attributed to host factor differences, particularly CD4, CD8 and Th17 activity,^{12 20} all of which may be affected by ruxolitinib.⁴ Additional factors associated with developing granulomatous diseases include exposure time for Pneumocystis glycoproteins, absence of IgA Pneumocystis antibodies and prior Pneumocystis exposure. Genotypic variation of Pneumocystis, once thought to play a role, is now considered less likely to contribute to the development of granulomas.²¹

Diagnostics

Physicians managing immunocompromised patients, such as those on ruxolitinib, must be aware of the less common nodular or cavitary PJP presentations to avoid delayed or missed diagnoses.¹ Cavitary lesions caused by tuberculosis may present similarly,²² but if acid-fast staining and mycobacterial cultures are negative, PJP should be considered as a cause of the cavitations. Coinfections with tuberculosis and PJP may also occur.^{23 24} Notably, as observed in our case, PTH-independent hypercalcaemia is associated with granulomatous PJP infections and should raise suspicion for this diagnosis.²⁵,²⁷

β-D-glucan is a helpful first test to check when granulomatous PJP is suspected and may adequately rule out a diagnosis of PJP if the pretest probability is low.²⁸ Lactate dehydrogenase has traditionally been used as a biomarker for PJP but a meta-analysis suggests that it has a poor sensitivity and specificity for the diagnosis of PJP overall.²⁹ Specificity of lactate dehydrogenase is expected to be particularly low in patients with high cell turnover from haematologic malignancies.¹²

A moderate-to-high pretest probability warrants checking for PJP in sputum or bronchoalveolar lavage specimens. Microscopic examination of sputum or lavage specimens may be sufficient to detect PJP yeast forms and make a diagnosis. However, the concentration of PJP tends to be lower in the sputum of non-HIV patients,³⁰ likely reducing the sensitivity of microscopy even with efforts to induce sputum to increase the sample volume.²⁹,³¹ PCR tests may therefore be preferred as a more sensitive test compared with microscopy, even if immunofluorescence assays are used for microscopic examination.³² However, distinguishing between active disease and colonisation in immunocompromised patients may require quantitative PCR in some clinical circumstances.³³

If sputum or lavage specimen analysis is inconclusive, transbronchial biopsy, transthoracic biopsy or even larger resections via thoracoscopic procedures, as in this case, may be needed for a definitive diagnosis.¹² Although larger studies are needed, thoracic surgery appears generally safe in immunocompromised patients, with the benefits of establishing a definitive diagnosis likely outweighing the risks of surgery in many cases.³⁴ However, a thorough discussion of the patient’s goals of care and prognosis from underlying immunocompromising conditions is of the utmost importance before pursuing such invasive procedures. Importantly, as demonstrated by Zhang et al,³⁵ some cases of cavitary PJP lesions can be diagnosed without resorting to thoracoscopic biopsy, underscoring the importance of exhausting less invasive diagnostic approaches when clinically appropriate.

Treatment & Prophylaxis

Current guidelines for managing PJP in immunosuppressed patients on immunotherapeutic and molecular-targeted agents do not provide specific treatment algorithms.⁴ Although TMP-SMX is considered first-line therapy for PJP based on randomised trials in HIV-infected patients,³⁶ alternative agents such as atovaquone or clindamycin plus primaquine may be used if TMP-SMX is contraindicated. Limited data suggest that outcomes for granulomatous PJP treatment may be comparable to those for other PJP infections.¹²

While severe PJP infections typically require 21 days of treatment, the optimal duration for granulomatous or cavitary PJP remains uncertain due to potentially higher organism burden and reduced drug penetration. In our case, treatment was extended to multiple months based on the patient’s immunocompromised state, extensive lung involvement and slow radiographic improvement. Although this approach led to improvement, such prolonged courses are not routine. Zhang et al³⁵ reported successful treatment of cavitary PJP lesions with a standard 3-week course of TMP-SMX in a patient with common variable immune deficiency, suggesting that shorter courses may still be effective in some cases of granulomatous PJP infections.

The efficacy of steroids in preventing inflammatory damage in non-HIV-infected individuals with PJP remains uncertain.³⁷,³⁹ Additionally, the benefits of ruxolitinib discontinuation in PJP infection are uncertain, given worsening haematologic disease activity can itself cause immunocompromise. Further complicating risk–benefit decisions, if ruxolitinib is stopped, there is a need to monitor closely for ruxolitinib discontinuation syndrome.⁴⁰ There may be a benefit of rechallenging with ruxolitinib if it is temporarily discontinued to help resolve an infection.⁴¹

There are no specific guidelines for prophylaxis of PJP in patients on ruxolitinib. Green et al⁴² suggest that a PJP risk greater than 3.5% warrants prophylaxis in immunocompromised non-HIV-infected adults. Although the overall risk of PJP in patients with myelofibrosis taking ruxolitinib does not appear to reach this threshold,⁴ further studies are needed to understand the risk in subsets of patients with evidence of more severe immunosuppression such as prior opportunistic infections or severe cytopenias.

Patient’s perspective.

The fungal infection I have been experiencing has made a difference in my health, initially causing many symptoms that truly affected me, chills, loss of appetite, shortness of breath, coughing, night sweats, loss of energy. Once I was diagnosed and doctors found medication, I gradually improved. Thanks so much for the care I received.

Learning points.

• In immunocompromised patients presenting with nodular or cavitary lesions, clinicians should consider Pneumocystis jirovecii pneumonia (PJP) in their differential diagnosis.

• Parathyroid hormone-independent hypercalcaemia should prompt clinicians to consider granulomatous PJP in immunocompromised patients.

• If initial diagnostics fail to confirm suspected granulomatous PJP, clinicians should weigh repeating less-invasive biopsy techniques against proceeding to potentially higher-yield thoracoscopic biopsies, considering the urgency for a definitive diagnosis.

• The optimal duration of treatment for PJP infections causing necrotising granulomas warrants further investigation, with factors such as degree of immunosuppression and extent of infection potentially influencing the necessity for prolonged therapy.