Marantic Endocarditis Presenting With Ischemic Stroke in Advanced Non-small Cell Lung Cancer: The Contribution of Fluorodeoxyglucose (FDG) PET/CT to the Differential Diagnosis of Valvular Vegetations

Abstract

Non-bacterial thrombotic endocarditis (NBTE), also known as marantic endocarditis, is an uncommon complication of advanced malignancy and an important cause of systemic embolism. Differentiating NBTE from infective endocarditis (IE) can be challenging because echocardiographic findings may be similar, while microbiological investigations remain negative. We report the case of a 66-year-old man with metastatic non-small cell lung cancer (NSCLC) who presented with multifocal ischemic stroke. Transthoracic and transesophageal echocardiography identified an 8.5 mm vegetation on the posterior mitral leaflet. Repeated blood cultures and an extensive evaluation for culture-negative IE were negative. Fluorodeoxyglucose (FDG) PET/CT demonstrated progression of the underlying malignancy but showed no abnormal uptake at the level of the cardiac valves. When interpreted together with the clinical and microbiological findings, these results supported the diagnosis of NBTE after exclusion of IE. This case highlights the potential contribution of FDG PET/CT as part of a multimodal diagnostic approach for differentiating sterile from infective valvular vegetations in patients with advanced cancer.

Introduction

Non-bacterial thrombotic endocarditis (NBTE), previously known as marantic endocarditis, is characterized by the formation of sterile fibrin and platelet thrombi on the valve leaflets, usually along the valve closure line, in the absence of bacteremia [1]. Even though it was described more than a century ago [2], it remains a diagnostic challenge and is often missed in patients with advanced cancer. Autopsy series report NBTE in 0.9-1.6% of the general population and in up to 19% of cancer patients, with lung cancer being the most frequently implicated malignancy, accounting for 20-30% of cases [3,4].

The pathophysiology is rooted in the cancer-associated hypercoagulable state: tumor-derived tissue factor, mucin secretion by adenocarcinomas, and systemic endothelial injury collectively promote platelet-fibrin deposition on morphologically intact valve leaflets [5]. The resulting vegetations are typically small (2-10 mm), flat, non-destructive, and most commonly located on the mitral or aortic valve, yet they are associated with a high risk of embolization, particularly to the central nervous system.

The principal diagnostic challenge is distinguishing NBTE from infective endocarditis (IE). Echocardiographic appearances can overlap considerably, and the modified Duke criteria cannot, by definition, identify NBTE. The 2023 Duke-International Society Cardiovascular Infectious Diseases (ISCVID) criteria and the 2023 European Society of Cardiology (ESC) Guidelines for the Management of Infective Endocarditis represent the most current frameworks for IE classification and now formally incorporate fluorodeoxyglucose (FDG) PET/CT as a major imaging criterion, including for native valve endocarditis [6,7]. However, the absence of valvular FDG uptake cannot reliably exclude IE in isolation and must be interpreted within a multimodal clinical framework.

We report a case of NBTE in a patient with progressive metastatic NSCLC initially presenting as multifocal embolic stroke, in whom a structured multimodal workup - including transesophageal echocardiogram (TEE), repeated blood cultures, comprehensive serologies, and FDG PET/CT - led to the correct diagnosis, appropriate management, and timely integration of palliative care.

Case presentation

Patient history

A 66-year-old male current smoker was admitted to the neurovascular unit for management of an acute ischemic stroke. His oncologic history included metastatic non-squamous non-small cell lung cancer (NSCLC) (cT1bN2M1c), diagnosed 18 months prior and currently under second-line chemotherapy with weekly paclitaxel following progression on carboplatin-based induction with immunotherapy maintenance. He also had a history of bladder cancer treated by radical cystoprostatectomy with orthotopic neobladder reconstruction five years earlier. His regular medication included apixaban 5 mg twice daily, prescribed in the context of cancer-associated prothrombotic risk.

Neurological presentation

Brain MRI at admission demonstrated multiple acute ischemic lesions in bilateral fronto-parieto-temporal and left temporal cortical-subcortical distributions, with right M2 segment occlusion and cortical hyperperfusion on arterial spin labeling (ASL) sequences, consistent with a cardioembolic mechanism. A 24-hour follow-up MRI showed the appearance of several new ischemic spots in bilateral hemispheric distributions. Chronic thrombosis of the right internal jugular vein was also identified, consistent with the known prothrombotic cancer context.

Cardiac imaging

Transthoracic echocardiography (TTE, Figure 1) revealed preserved left ventricular systolic function (LVEF, 60%) and a 9 mm vibrating echogenic image on the atrial surface of the posterior mitral leaflet, highly suspicious for a vegetation. Trivial mitral regurgitation and moderate pulmonary hypertension (estimated pulmonary artery systolic pressure of 46 mmHg) were noted. No pericardial effusion was detected.

Figure 1. Apical four-chamber transthoracic echocardiography (TTE) image of our case demonstrating a hyperechogenic mobile vegetation measuring 8.4 mm (caliper measurement shown) on the atrial aspect of the posterior mitral leaflet.

The vegetation is identified at the level of the mitral valve coaptation line, without evidence of leaflet perforation or perivalvular extension.

Transesophageal echocardiography (TEE) (Figure 2) confirmed a filamentous vibrating image of 8.5 mm on the atrial aspect of the posterior mitral leaflet with mild central regurgitation, thickening of both leaflet free borders, and thickening of the mitro-aortic trigone without color flow. Left atrial appendage flow velocity was 70 cm/s with no spontaneous echo contrast and no thrombus. The aortic valve was thin and tricuspid with no additional elements and no regurgitation. Left ventricular morphology and kinetics were normal.

Figure 2. Multiplane transesophageal echocardiography (TEE) images of our case showing a filamentous, highly mobile vegetation on the atrial aspect of the posterior mitral leaflet (red arrows) in four imaging planes, with measurements ranging from 7.1 to 9.0 mm across the planes.

No perivalvular abscess, leaflet perforation, or prosthetic material involvement was identified.

Infectious workup

Diagnostic criteria for IE were systematically applied using the modified Duke criteria and, retrospectively, the 2023 Duke-ISCVID criteria and the 2023 ESC guidelines [6-8]. Three sets of blood cultures were obtained from peripheral venous sites on separate occasions, with aerobic and anaerobic incubation maintained for five days. All cultures were negative. Initial sets were drawn prior to any antibiotic administration; one subsequent set was obtained after initiation of antibiotherapy for a concurrent pulmonary infection, though all cultures - including those drawn before antibiotic exposure - were equally negative, providing robust microbiological evidence against bacteremia.

Serological screening included Bartonella henselae and B. quintana (IgG/IgM), Coxiella burnetii phase II IgG/IgM, and Brucella (Wright agglutination and Rose Bengal test), all showing no evidence of active infection. Serological testing for Mycoplasma pneumoniae and Chlamydia pneumoniae demonstrated isolated IgG positivity without IgM, consistent with past exposure and excluding acute infection. HIV serology was negative. Urinary antigen for Legionella pneumophila serogroup 1 was negative. Urine culture yielded Enterococcus-type cocci at under 10³ CFU/mL, interpreted as contamination in the context of concurrent antibiotherapy. The totality of this workup provided no microbiological evidence supporting IE under any of the applied diagnostic frameworks.

According to the 2023 Duke-ISCVID criteria, the patient fulfilled one major imaging criterion (mobile mitral vegetation on transthoracic and TTE) and one minor criterion (multifocal embolic strokes as vascular phenomena), formally classifying the case as possible IE. However, persistently negative blood cultures across multiple independent sets, a comprehensive negative serological panel, absence of valvular FDG uptake on PET/CT, and the clinical context of advanced metastatic malignancy collectively favored a sterile thrombotic etiology, supporting the diagnosis of NBTE over infective endocarditis.

FDG PET/CT

FDG PET/CT (performed on a BIOGRAPH Vision 600, Siemens; 105 MBq 18F-FDG; fasting blood glucose: 5.2 mmol/L) demonstrated a hypermetabolic right paracardiac mass (SUVmax 26.4) and extensive hypermetabolic lymphadenopathy involving cervical, mediastinal (SUVmax 21.2 at the subcarinal level), and retroperitoneal stations, consistent with progression of the known NSCLC (Figure 3). Bilateral basal pulmonary consolidations and small bilateral pleural effusions were also present. Despite widespread hypermetabolic disease elsewhere, no abnormal FDG uptake was observed at the level of the cardiac valves or myocardium. Although the absence of valvular FDG uptake does not exclude infective endocarditis and is not, by itself, sufficient for diagnostic classification in native-valve disease, it provided supportive evidence against an active infectious process when interpreted alongside persistently negative blood cultures, negative serological investigations, and the clinical context. The marked discrepancy between intense metabolic activity related to metastatic malignancy and the absence of valvular uptake further supported a non-infectious vegetation consistent with non-bacterial thrombotic endocarditis (NBTE) [9].

Figure 3. Axial fused FDG PET/CT image of our case at the thoracic level demonstrating intense hypermetabolic activity in the right paracardiac region (SUVmax, 26.4) and bilateral lung bases, consistent with NSCLC progression.

Despite widespread hypermetabolic disease in the surrounding structures, no abnormal FDG uptake is observed at the level of the cardiac valves or within the myocardium, supporting the diagnosis of non-bacterial thrombotic endocarditis (NBTE).

Laboratory data

Laboratory investigations showed markedly elevated inflammatory markers throughout the hospitalization, with C-reactive protein ranging from 165 to 271 mg/L and a peak leukocyte count of 19.30 x 10³/mm³ with predominant neutrophilia (peak 87.0%). Procalcitonin was measured at 0.507 ng/mL, exceeding the laboratory reference threshold of 0.250 ng/mL. However, this mild elevation must be interpreted with caution in the context of advanced malignancy, progressive renal dysfunction, and concurrent pulmonary infection, as detailed in the Discussion. A normocytic anemia was present with a nadir hemoglobin of 7.0 g/dL, consistent with anemia of chronic disease in the setting of advanced cancer. Progressive renal dysfunction (peak creatinine: 255 µmol/L) was attributed to bilateral ureteral obstruction by retroperitoneal metastatic lymphadenopathy. Key laboratory findings, with admission, peak, and nadir values as appropriate, are summarized in Table 1.

Table 1. Key laboratory results during hospitalization.

Parameter	Patient Value	Reference Range	Unit
Biochemistry
Sodium	143	136-145	mmol/L
Potassium	4.5	3.4-4.5	mmol/L
Chloride	108 ↑	98-107	mmol/L
Urea (peak)	24.2 ↑	2.8-8.1	mmol/L
Creatinine (peak)	255 ↑	59-104	µmol/L
Total Proteins	63 ↓	66-87	g/L
Glucose	8.0 ↑	4.1-6.0	mmol/L
Calcium	2.23	2.20-2.55	mmol/L
Total Bilirubin	11	Feb-21	µmol/L
Lactate	1.5	0.5-2.2	mmol/L
Liver Enzymes
ASAT/GOT	27	<40	U/L
ALAT/GPT	9	<41	U/L
GGT	83 ↑	<60	U/L
ALP	106	40-130	U/L
Inflammatory & cardiac markers
CRP (range)	165-271 ↑	<5	mg/L
Procalcitonin (peak)	0.507 ↑	<0.250	ng/mL
Troponin T hs (peak)	39.5 ↑	<14	pg/mL
NT-proBNP	696	<2346	pg/mL
CPK	15	<190	U/L
Hematology
Hemoglobin	7.0 ↓	13.0-17.0	g/dL
Hematocrit	24.6 ↓	40-50	%
MCV	94	82-97	fL
WBC (peak)	19.30 ↑	04-Oct	×10³/mm³
Neutrophils %	87.0 ↑	52-68	%
Lymphocytes %	3.1 ↓	26-38	%
Platelets	225	150-450	×10³/mm³

Multidisciplinary decision and outcome

Following multidisciplinary discussion involving cardiology, cardiac surgery, infectious disease, oncology, and palliative care specialists, NBTE was considered the most likely diagnosis in the setting of progressive metastatic NSCLC. The absence of bacteremia across three independent culture sets, the negative serological panel, the absence of valvular FDG uptake on PET/CT, and the clinical context of advanced malignancy collectively supported this conclusion. Surgical intervention was unanimously contraindicated given the advanced oncological stage and unfavorable risk-benefit profile. Apixaban was switched to therapeutic low-molecular-weight heparin. In view of the patient's clinical deterioration and the palliative trajectory of his underlying malignancy, care goals were comprehensively reassessed in collaboration with the patient's family, and a palliative care framework was established.

Verbal informed consent was obtained from the patient's family for the publication of this case report and accompanying images, in accordance with institutional requirements.

Discussion

This case illustrates the diagnostic and therapeutic challenges associated with NBTE in a patient with progressive malignancy. It highlights the potential contribution of FDG PET/CT in the differential diagnosis of valvular vegetations, the importance of an exhaustive microbiological evaluation to exclude infective endocarditis, and the value of multidisciplinary decision-making in guiding patient management.

Pathophysiology

NBTE is a recognized complication of advanced malignancy and is largely a consequence of the hypercoagulable state associated with cancer. Non-squamous lung cancers upregulate tissue factor expression, stimulate platelet activation, and trigger chronic low-grade endothelial injury, creating a permissive environment for sterile thrombus formation at the line of valve closure [3,5]. The resulting vegetations are pathologically distinct from IE: they are acellular, non-inflammatory, and lack the destructive potential of infected vegetations. This acellular, non-inflammatory nature accounts for the absence of FDG uptake observed in our patient and is consistent with the absence of a major imaging criterion under the 2023 Duke-ISCVID criteria. Of note, the immunotherapy received by our patient has been associated with an increased prothrombotic risk, further supporting the pathophysiological substrate for NBTE in this case.

Diagnostic framework: modified Duke, Duke-ISCVID 2023, and ESC 2023 criteria

The diagnostic classification of IE has evolved significantly with the 2023 Duke-ISCVID criteria and the 2023 ESC Guidelines [6,7]. Compared with the 2000 modified Duke criteria and the 2015 ESC criteria [8], both updated frameworks expand the major imaging criteria to include FDG PET/CT findings in native valve endocarditis: under the 2023 Duke-ISCVID criteria, any abnormal FDG uptake pattern on native valve tissue now constitutes a major imaging criterion, whereas the 2015 ESC criteria restricted this role to prosthetic valve endocarditis. The 2023 ESC criteria similarly extend the role of FDG PET/CT for native valve disease. Both updated frameworks maintain the diagnostic thresholds of definite IE (2 major criteria, or 1 major and 3 minor criteria), possible IE (1 major criterion, or 3 minor criteria), and rejected IE (criteria not met for definite or possible IE).

Application of the 2023 Duke-ISCVID criteria to our patient identified one major imaging criterion (mobile mitral vegetation on echocardiography) and one minor criterion (multifocal embolic strokes as vascular phenomena), formally classifying the case as possible IE. No major microbiological criterion was fulfilled despite three independent blood culture sets and an extensive serological panel. Although the echocardiographic appearance initially raised concern for infective endocarditis, the absence of microbiological evidence, lack of valvular FDG uptake on PET/CT, and the presence of advanced metastatic NSCLC strongly supported a diagnosis of NBTE. This case illustrates that, even in the era of expanded multimodality imaging criteria, the diagnosis of NBTE continues to rely on careful integration of clinical, microbiological, echocardiographic, and multimodality imaging findings rather than on isolated imaging abnormalities.

Role of FDG PET/CT

FDG PET/CT has become an increasingly recognized adjunct in the diagnostic workup of suspected infective endocarditis. Since its incorporation into the 2015 ESC guidelines as a major imaging criterion for prosthetic valve infective endocarditis, its role has been further expanded in the 2023 Duke-ISCVID criteria and the 2023 ESC guidelines to include the evaluation of native-valve infective endocarditis [6,7]. In our case, the complete absence of metabolic uptake at the mitral valve, despite extensive FDG-avid metastatic disease elsewhere - including a hypermetabolic right paracardiac mass (SUVmax 26.4) - was particularly informative. Within the contemporary multimodality diagnostic framework, abnormal valvular FDG uptake has assumed an increasingly important role in the evaluation of suspected native-valve infective endocarditis. Although the absence of valvular FDG uptake cannot independently exclude infective endocarditis, it provided important complementary evidence against an infectious etiology when interpreted alongside persistently negative blood cultures, negative serological investigations, and the echocardiographic findings. This observation is consistent with the acellular and non-inflammatory nature of NBTE vegetations, which typically demonstrate minimal or absent FDG uptake, in contrast to the inflammatory activity commonly observed in infective endocarditis [9,10]. Collectively, these findings helped avoid unnecessary prolonged antibiotic therapy, reinforced the diagnosis of malignancy-associated NBTE, and supported the therapeutic strategy of anticoagulation.

Microbiological workup

Three sets of blood cultures were obtained from peripheral venous sites on separate occasions, with aerobic and anaerobic incubation maintained for five days. All cultures were negative. Initial sets were drawn prior to any antibiotic administration; one subsequent set was obtained after initiation of antibiotherapy for a concurrent pulmonary infection, though all cultures - including those drawn before antibiotic exposure - were equally negative, providing robust microbiological evidence against bacteremia. The exhaustive serological panel covering the most common fastidious organisms implicated in culture-negative IE - B. henselae and B. quintana, C. burnetii, and Brucella - was similarly unrevealing. For M. pneumoniae and C. pneumoniae, isolated IgG positivity without IgM was consistent with remote exposure only. The combination of repeatedly negative cultures across independent sampling occasions and a comprehensive serological screen, together with the absence of valve-level FDG avidity, provided convergent evidence supporting a non-infectious etiology, with no major microbiological criterion fulfilled under any of the applied diagnostic frameworks.

Interpretation of inflammatory markers

The inflammatory profile in this case - markedly elevated CRP (165-271 mg/L), leukocytosis with neutrophilia (peak: 19.30 x 10³/mm³, 87.0%), and mildly elevated procalcitonin (0.507 ng/mL, exceeding the reference threshold of 0.250 ng/mL) - warrants careful interpretation. While these findings may superficially suggest an infectious process, they must be contextualized within the clinical setting of advanced metastatic NSCLC, progressive renal dysfunction secondary to ureteral obstruction, and a concurrent pulmonary infection requiring antibiotherapy. Each of these conditions is independently capable of driving significant elevation of CRP, leukocytosis, and procalcitonin, none of which are specific markers for infective endocarditis. The mildly elevated procalcitonin, while above the laboratory reference range, remained well below the thresholds associated with severe bacterial infection (typically greater than 2 ng/mL for sepsis and greater than 10 ng/mL for septic shock), and its elevation is most plausibly attributed to the concurrent pulmonary infection and renal impairment rather than to valvular infection. The aggregate biomarker profile therefore does not support a diagnosis of bacterial endocarditis and must be interpreted within the full multimodal clinical framework.

Anticoagulation

Therapeutic anticoagulation represents the cornerstone of NBTE management. Rogers et al. demonstrated that heparin-based anticoagulation reduces embolic recurrence without significantly increasing hemorrhagic risk [11]. In oncologic patients, low-molecular-weight heparin (LMWH) is preferred over vitamin K antagonists, consistent with American Society of Hematology 2021 guidelines for cancer-associated thromboembolism, given its demonstrated superiority in reducing recurrent thromboembolism. It is acknowledged, however, that LMWH undergoes renal clearance and that dose adjustment with anti-Xa monitoring is warranted in the context of significant renal dysfunction, as observed in our patient [12]. In our patient, pre-admission apixaban therapy prescribed for cancer-associated venous thromboembolism failed to prevent the embolic event, prompting a switch to therapeutic LMWH after the diagnosis of NBTE was established. The failure of apixaban to prevent embolization may reflect the platelet-rich, thrombin-independent nature of NBTE vegetations, which are less responsive to direct oral anticoagulants targeting the coagulation cascade; however, this mechanistic explanation remains speculative and should be regarded as hypothesis-generating rather than evidence-based, pending prospective evaluation.

Surgical considerations

Surgery for NBTE in the setting of advanced malignancy is almost universally contraindicated. Without effective treatment of the primary tumor, sterile vegetations will inevitably reform on any surgically repaired or replaced valve. Eiken et al., in a Mayo Clinic series of 30 surgical NBTE cases, demonstrated that the underlying hypercoagulable state - not valvular anatomy - drives prognosis [13]. In our case, the unanimous multidisciplinary decision to forego surgery was well-supported by both the evidence base and the patient's advanced clinical trajectory, including progressive oncological disease and clinical deterioration requiring palliative care integration.

Prognostic implications and palliative care integration

In the present case, NBTE occurred in the setting of progressive metastatic NSCLC on second-line chemotherapy, with evidence of disease progression on FDG PET/CT and rapid clinical deterioration. Beyond its thromboembolic consequences, the diagnosis of NBTE in this context reflects the advanced and refractory nature of the underlying malignancy and should prompt immediate reassessment of therapeutic goals. Early and proactive integration of palliative care - encompassing pain management, goals-of-care discussions, and family support - is essential and was undertaken in the present case following multidisciplinary review. This case reinforces that, in patients with advanced cancer and NBTE, the primary determinant of prognosis and the primary target of management is the underlying malignancy itself, not the valvular lesion.

Conclusions

NBTE should be considered in patients with active malignancy presenting with embolic stroke and valvular vegetations. Diagnosis relies on careful integration of echocardiographic findings, repeated microbiological investigations, and clinical context. Application of the 2023 Duke-ISCVID criteria and the 2023 ESC guidelines provides the most current and sensitive diagnostic framework: in our case, both frameworks classified the case as possible IE, reflecting the presence of one major imaging criterion (mitral vegetation) and one minor criterion (embolic strokes); however, the absence of any microbiological evidence and the overall multimodal assessment strongly favored NBTE over infective endocarditis. Inflammatory biomarkers, including CRP, leukocytosis, and mildly elevated procalcitonin, must be interpreted with caution in the context of advanced malignancy and concurrent infection, as they lack specificity for bacterial endocarditis in this setting. FDG PET/CT may provide useful complementary information within a multimodal approach; under the 2023 criteria, the complete absence of valvular FDG uptake, while not independently excluding IE, constitutes the absence of a major imaging criterion and contributes meaningfully to the overall diagnostic assessment. Management is centered on therapeutic anticoagulation, treatment of the underlying disease, and early palliative care integration when appropriate.