CMR Algorithm Using Dark-Blood LGE and Parametric Mapping Accurately Differentiates Reversible Inflammatory From Irreversible Fibrotic Constriction: Outcomes-Anchored Study in a TB-Endemic Population

Background: Distinguishing inflammatory pericarditis with constrictive physiology (potentially reversible) from chronic fibrotic constriction (surgical) remains a clinical challenge. Cardiac MRI (CMR) markers—including pericardial late gadolinium enhancement (LGE), T1/T2 mapping, and real-time cine features—may identify patients likely to respond to medical therapy. Dark-blood LGE further enhances pericardial signal conspicuity, but data on outcome-anchored validation are limited, particularly in tuberculosis (TB)-prevalent regions.

Methods: We retrospectively studied 72 consecutive adults with constrictive physiology confirmed by echocardiography and/or catheterization who underwent CMR between 2020–2024. The standardized protocol included cine imaging, real-time free-breathing sequences, dark-blood phase-sensitive inversion recovery (PSIR) LGE, and native T1/T2 mapping.
Patients were categorized at baseline CMR as:

1. Inflammatory (pericardial LGE ± elevated T2, septal bounce),


2. Fibrotic (pericardial thickening without LGE or T2 elevation),


3. Mixed/indeterminate (patchy features).

Clinical adjudication at 3–6 months served as the reference: resolution on medical therapy (NSAIDs/steroids ± anti-TB) vs. persistent constriction requiring pericardiectomy.

Results: Inflammatory phenotype was identified in 29/72 (40%), fibrotic in 31 (43%), and mixed in 12 (17%). Outcomes are summarized in Table 1.

• Inflammatory pattern predicted clinical resolution in 76% (22/29).


• Fibrotic pattern was associated with persistent constriction or surgery in 87% (27/31).


• Mixed group showed intermediate outcomes (58% improved).

Diagnostic performance: inflammatory pattern predicted reversibility with sensitivity 79%, specificity 85%, AUC 0.86. Addition of LGE/T2 to pericardial thickness reclassified 22% of cases correctly (NRI, p=0.01). Dark-blood LGE improved detection of pericardial enhancement vs. conventional bright-blood (p< 0.01) and enhanced reader confidence. Inter-observer agreement for pattern assignment was high (κ=0.86).

Conclusion: A simple CMR algorithm incorporating dark-blood LGE, T1/T2 mapping, and cine signs accurately separates reversible inflammatory constriction from fixed fibrosis. This approach can directly guide therapy, reducing unnecessary surgery while expediting pericardiectomy when irreversible disease is present. Outcomes-anchored validation in a TB-endemic cohort underscores its global relevance.

Table 1. Baseline CMR Patterns and 3–6 Month Clinical Outcomes (Nf72).

CMR Classification	n (%)	Key Imaging Features	Resolution Rate	Persistent/Surgery Rate
Inflammatory	29 (40%)	Pericardial LGE, ↑ T2, septal bounce	22 (76%)	7 (24%)
Fibrotic	31 (43%)	Thickened pericardium, no LGE or T2 elevation	4 (13%)	27 (87%)
Mixed / Indeterminate	12 (17%)	Patchy LGE ± thickening	7 (58%)	5 (42%)

Table 2. Incremental Diagnostic Value of CMR Algorithm Beyond Pericardial Thickness.

Model	AUC (95% CI)	NRI (%)	Key Added Value	p-value
Pericardial thickness alone	0.71 (0.60–0.82)	—	Limited discrimination	Reference
+ Dark-blood LGE	0.83 (0.74–0.92)	+14%	Enhanced detection of inflammation	0.01
+ Dark-blood LGE + T1/T2 mapping	0.86 (0.78–0.94)	+22%	Best classification of reversibility	< 0.01

Figure 1: Outcomes at 3-6 Months by CMR Pattern.