Contrast-enhanced computed tomography brain as a cost-effective imaging tool for guiding mechanical thrombectomy in resource-limited settings

INTRODUCTION

Mechanical thrombectomy has emerged as a life-saving and disability-reducing intervention for patients with acute ischemic stroke caused by large vessel occlusion (LVO).^[1] It is now established as the standard of care for eligible patients, offering significantly better outcomes compared to medical therapy alone, particularly when performed within the recommended therapeutic window. Timely restoration of cerebral perfusion through thrombectomy can result in improved neurological recovery, reduced long-term disability, and enhanced quality of life.

Despite its proven benefits, access to mechanical thrombectomy remains limited in many low- and middle-income countries (LMICs), particularly in rural areas and underserved urban regions.^[2] A major barrier is the timely and accurate diagnosis of LVO, which is essential for selecting appropriate candidates for intervention. In the absence of advanced imaging modalities, stroke management often defaults to medical therapy based on clinical assessment alone.

Before the introduction of mechanical thrombectomy services at our center, stroke evaluation relied solely on non-contrast computed tomography (CT) brain imaging. Once intracranial hemorrhage was excluded, management decisions were guided by the time since symptom onset and clinical presentation, often resulting in conservative treatment. With the establishment of a dedicated thrombectomy program, the need for improved LVO detection became evident.

Although CT angiography (CTA) remains the gold standard for non-invasive detection of LVO, its application is frequently limited by cost and resource constraints. In our setting, CTA costs approximately INR 12,000 (141.83$)—rendering it inaccessible to many patients. In contrast, contrast-enhanced CT (CECT) of the brain is significantly more affordable at INR 2,000 (23.64$) and is more widely available.

In response to these limitations, we developed a pragmatic, resource-sensitive protocol that utilizes CECT brain with manual contrast injection for LVO detection. While CECT does not match CTA in terms of spatial resolution and vascular detail, it offers a viable alternative for timely triage and treatment planning in acute ischemic stroke. In this study, we assessed the diagnostic performance of CECT brain in detecting LVO and compared it with established reference standards, including digital subtraction angiography (DSA), CTA, and magnetic resonance angiography (MRA). Our goal was to evaluate whether this simplified imaging strategy could effectively guide patient selection for mechanical thrombectomy in a resource-limited environment.

MATERIALS AND METHODS

This was a retrospective observational study. Ethical committee approval was obtained, and informed consent was taken. This study included all patients who underwent mechanical thrombectomy at our institute, where the decision for intervention was based on findings from CECT brain imaging.

CECT scans were performed following intravenous administration of 20 mL of iodinated contrast agent, with image acquisition initiated approximately 15–16 s after injection onset. Axial images were reconstructed at 2 mm slice thickness and evaluated for evidence of LVO. Occlusions involving the proximal or distal internal carotid artery, M1 segment of the middle cerebral artery (MCA), proximal M2 branches, intracranial vertebral arteries, or basilar artery were categorized as LVO.

Patients presenting with acute ischemic stroke who were deemed eligible for thrombectomy within a 24-hour window were taken up for the procedure. Patient selection for thrombectomy was based either on definite LVO on CECT or strong clinical-imaging mismatch, where LVO was strongly suspected. In all cases, LVO was confirmed on DSA immediately before thrombectomy.

In addition, we evaluated a control cohort of patients (n = 25) who presented with clinical features suggestive of ischemic stroke and underwent CECT brain imaging. These patients demonstrated no evidence of LVO on CECT and subsequently underwent confirmatory imaging with CTA or MRA within 24 h, which verified the absence of LVO. This control group was included to support the calculation of diagnostic performance measures, including sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) for CECT in LVO detection.

Functional outcomes were assessed using the modified Rankin Scale (mRS) at discharge. Scores of 0–2 were categorized as “good outcomes,” while scores of 3–6 were defined as “poor outcomes.”

Diagnostic performance metrics for CECT were calculated using standard 2 × 2 contingency tables. Correlations between imaging findings, thrombectomy success, and clinical outcomes were assessed using Spearman’s rank-order correlation tests. Statistical significance was set at P < 0.05.

RESULTS

A total of 25 patients underwent mechanical thrombectomy based on LVO identified on CECT brain imaging. The mean age of the cohort was 48.6 years, and the group included 19 males and 6 females. The age group distribution is mentioned in Table 1. All procedures utilized aspiration techniques for recanalization. The sites of occlusions are mentioned in Table 2. Table 3 shows 2 × 2 contingency table for the calculation of sensitivity, specificity, PPV and NPV.

CECT demonstrated excellent diagnostic performance in detecting LVO. The sensitivity was 96.0% (95% CI: 79.65–99.9%), and specificity was 100% (95% CI: 86.2–100%). The PPV was 100% (95% CI: 85.75–100%), and the NPV was 96.15% (95% CI: 78.56–99.42%). These results indicate that a positive CECT reliably confirms LVO, while a negative result is strongly suggestive of its absence, although a small risk of false negatives remains. Representative CECT brain images are shown in Figure 1. Diagnostic performance metrics are illustrated in Figure 2.

Close

Figure 1:

(a and b) contrast-enhanced computed tomography (CECT) images of patient with left middle cerebral artery (MCA) occlusion (arrow in a), with normal filling of contralateral MCA (arrow in b), (c and d) CECT axial images of patient with terminal internal carotid artery (ICA) occlusion (straight arrow in c shows occlusion of left terminal ICA; curved arrow in c shows patent right terminal ICA). Proximal sections (d) show patent left proximal ICA (straight arrow) and patent right proximal ICA (arrowhead) (e and f) CECT axial images of patient with proximal + terminal ICA occlusion. Proximal sections (e) show patent right proximal ICA (arrowhead) and occluded left proximal ICA (straight arrow). CECT brain at upper level (f) shows occluded left terminal ICA (straight arrow) and patent right terminal ICA (curved arrow). (g and h) CECT axial images of patient with false negative finding on CECT brain shows no obvious occlusion of left M1 segment- arrows in g and h (however this patient had M1 occlusion on DSA).

Export to PPT

Close

Figure 2:

Diagnostic performance of contrast-enhanced computed tomography for large vessel occlusion detection. Error bars represent the 95% confidence intervals for each metric. PPV:Positive predictive value, NPV: Negative predictive value.

Export to PPT

The case, which was a false negative on CECT, was reviewed retrospectively. The image was acquired after multiple attempts due to multiple motion artifacts in previously acquired images; hence, there was likely delayed enhancement of the clot and decreased density of contrast in vessels. Furthermore, there was a hyperdense left MCA M1 segment. All these findings likely led to false-negative results. Based on high National Institutes of Health Stroke Scale (NIHSS), which showed deterioration along with strong clinical suspicion of LVO, along with the presence of hyperdense MCA sign, the patient was taken up for thrombectomy.

Of the 25 patients, 17 (68%) achieved good functional outcomes (mRS 0–2), while 5 patients (20%) experienced fatal outcomes (mRS 6). Three patients (12%) experienced post-thrombectomy hemorrhage. Among them, one patient recovered well (mRS 1), one had moderate disability (mRS 4), and one died. These findings highlight both the potential benefits and risks of the procedure in this setting.

Recanalization success, defined as modified thrombolysis in cerebral infarction score ≥2b, was significantly associated with favorable outcomes. A Spearman rank-order correlation revealed a strong inverse association between thrombolysis in cerebral infarction (TICI) score and mRS outcome (Spearman’s ρ = –0.61; P = 0.001), suggesting that complete or near-complete reperfusion is linked to better recovery.

Pre-procedure Alberta Stroke Program Early CT Score (ASPECTS) was also predictive of the outcome. A significant negative correlation was observed between initial ASPECTS and mRS scores (Spearman’s ρ = –0.56; P = 0.002), indicating that higher ASPECTS values were associated with better clinical outcomes. This is depicted in Figure 3.

Close

Figure 3:

Boxplot comparing pre-procedure Alberta stroke program early computed tomography scores (ASPECTS) between patients with good and poor outcomes. mRS: Modified rankin scale.

Export to PPT

DISCUSSION

In standard clinical practice, CTA is considered the modality of choice for detecting LVO due to its high sensitivity and specificity, with reported sensitivity around 93% and specificity reaching 100% in acute ischemic stroke settings.^[3] However, in resource-constrained environments, the routine use of CTA poses several practical challenges. It requires trained technical staff, a contrast pressure injector, and administration of a high volume of iodinated contrast agent—typically 80 mL per scan.^[4] Moreover, the cost of CTA is substantially higher – approximately 6 times that of a CECT brain study – posing a significant barrier to access for patients in LMICs.

At our institution, these limitations became evident after the initiation of mechanical thrombectomy services. Financial constraints and limited infrastructure often precluded the use of CTA, leading to missed diagnoses of LVO and delayed or missed opportunities for intervention in otherwise eligible patients. To address this gap, we adopted a protocol utilizing CECT brain imaging to identify LVO. Although this method is not intended to replace CTA, it serves as a pragmatic and cost-effective alternative in resource-limited settings. Our findings support the feasibility of using CECT-based protocols to expand access to endovascular stroke treatment in such environments, particularly where CTA is not routinely available.

In this study, CECT demonstrated exceptional diagnostic performance in detecting LVO, with a sensitivity of 96% and specificity of 100% compared to the gold standard of DSA, CTA, or MRA. The high PPV (100%) further reinforces the reliability of a positive CECT finding in confirming LVO, while the NPV of 96.2% indicates a strong likelihood of ruling out LVO when the scan appears normal. These results support the use of CECT as a viable alternative imaging tool for identifying candidates for mechanical thrombectomy, especially in emergent settings where time-sensitive decisions must be made.

Although standard CECT brain protocols typically use 50– 80 mL of iodinated contrast, we employed a reduced volume of 20 mL based on consistent observations of adequate intracranial arterial opacification for LVO assessment using this volume of contrast with manual injection. This protocol was developed in response to practical challenges in our setting, including limited availability of contrast agents, lack of access to pressure injectors, and the need to minimize procedural cost, as patients are required to directly bear the expenses of imaging. Given these constraints, manual injection of a lower contrast volume offered a feasible, cost-effective alternative while still providing sufficient diagnostic information for clinical decision-making in suspected LVO.

Duvekot et al. (2021) emphasized the importance of operator experience and real-world variability in interpreting CTA results, and that its accuracy can be affected by clinical setting and reader expertise.^[3] Similar considerations apply to CECT brain, which should be interpreted in conjunction with clinical findings such as a high NIHSS score. In addition, CECT brain may also aid in the calculation of core infarct volume, similar to CT angiography source images. A CECT brain can also rule out stroke mimics such as infection, tumor, and cerebral venous thrombosis.

Nevertheless, the CECT brain is subject to various limitations. These include the possibility of venous contamination, particularly if imaging acquisition is delayed or due to variable arm-to-brain circulation time. As mentioned previously, there may be delayed enhancement of the clot and decreased intravascular density of contrast in cases of delayed acquisition, which may lead to erroneous interpretation. Furthermore, collateral circulation cannot be reliably assessed with CECT due to its single time-point acquisition, for which multiphase CTA is considered superior.^[5] Importantly, CECT does not provide information regarding the cervical vessels, aortic arch configuration, vascular tortuosity, or the presence of carotid web or plaque, all of which are assessable with CTA, thereby affirming the latter’s diagnostic superiority.^[6]

The lack of access to brain imaging remains a significant barrier to optimal stroke management in LMICs.^[2] A systematic review from Africa reported that only 13–36% of stroke patients underwent brain imaging, even in regions equipped with functional CT or MRI machines.^[7] Similarly, a study from rural India found that only 12% of patients received any form of brain imaging.^[8] In such settings, a simplified imaging workflow may play a pivotal role in the early identification of LVO and timely referral to centers equipped for mechanical thrombectomy.

Cost-effectiveness analyses consistently show that comprehensive imaging – typically involving a non-contrast CT followed by CTA and/or CT perfusion – yields the highest lifetime quality-adjusted life years and represents the most economically favorable strategy when accounting for the benefits of thrombectomy.^[9] However, these advanced imaging modalities are largely unavailable or unaffordable across most regions in India. An alternative model, the “direct transfer to angiography suite” (DTAS) approach, has demonstrated both improved outcomes and reduced costs for LVO patients presenting within 6 h of symptom onset, when compared to the conventional “direct to CT” workflow.^[10] Yet, implementation of DTAS in LMICs is largely impractical presently, due to infrastructure limitations. Even in centers where angiography suites are available—such as ours—there remains a shortage of essential trained personnel, including an adequate number of neurointerventionalists, neuroanesthetists, specialized nursing staff, and cath-lab technicians, which hinders round-the-clock availability for immediate angiography in all suspected LVO cases.

Earlier initiation of treatment following stroke onset is well established as a key determinant of improved functional outcomes and recovery.^[11-13] In this context, even modest enhancements to the pre-treatment imaging pathway – such as performing a CECT brain scan when CTA is unavailable, unaffordable, or logistically not feasible – can significantly influence clinical outcomes in acute ischemic stroke patients.

In our cohort of 25 patients undergoing mechanical thrombectomy, 68% (17/25) achieved a good clinical outcome (mRS 0–2), and 20% (5/25) had a fatal outcome (mRS 6). These rates are comparable to those observed in major randomized trials. For instance, the MR CLEAN trial reported 32.6% good outcome in the intervention group, while EXTEND-IA and SWIFT PRIME studies showed significantly higher good outcome rates of 71% and 60%, respectively.^[14-16] Our results, especially the good outcome rate, align more closely with these latter trials, suggesting effective patient selection and procedural success.

The TICI 3 recanalization rate in our study was 68% (17/25), which compares favorably with published literature. The HERMES meta-analysis, which pooled data from five landmark thrombectomy trials, reported a TICI 2b/3 rate of approximately 71%.^[1] In addition, the strong statistical association we observed between TICI 3 and good outcome (P = 0.00013; OR = 112) reinforces existing evidence that complete recanalization is a key predictor of functional recovery.

Moreover, pre-procedure ASPECTS scores showed a trend toward predicting favorable outcomes, reinforcing the role of imaging not just in diagnosis but also in prognostication. This further underlines the importance of early and effective imaging in stroke care, even if only a CECT is available initially.

In our study, post-thrombectomy hemorrhage occurred in 8% (2/25) of patients, with outcomes ranging from full recovery to death, highlighting the variable clinical impact of this complication. Although our hemorrhage rate is slightly higher than the 4.4% reported in the HERMES meta-analysis, the absence of other major adverse events such as vessel perforation or embolization supports the overall safety of the procedure.^[1] These results underscore that with careful technique and patient selection, mechanical thrombectomy remains a safe intervention, even in resource-limited settings.

Given the favorable diagnostic accuracy, lower cost, and broader accessibility of CECT, it may serve as an effective frontline modality for triaging patients for mechanical thrombectomy in resource-constrained environments. While CTA remains the gold standard, CECT can offer a pragmatic compromise when CTA is not immediately available, helping to prevent delays in treatment. Training emergency physicians and radiologists in interpreting CECT for signs of LVO could expand its utility and ensure more equitable access to potentially life-saving stroke interventions such as thrombectomy. Our findings support that even with simplified imaging workflows using CECT, effective thrombectomy outcomes can be achieved, thus widening access to life-saving stroke therapy in under-resourced healthcare systems.

CONCLUSION

This study demonstrates that mechanical thrombectomy, when guided by contrast-enhanced CT, can achieve clinical outcomes and recanalization success rates comparable to those reported in high-resource settings. CECT proved to be a highly accurate, cost-effective, and accessible imaging tool for identifying LVO, making it especially valuable in resource-limited environments. CECT offers several practical advantages over CTA, including lower cost, wider availability, and simpler technical requirements. With CTA costing nearly 6 times more than CECT in our setting and requiring a pressure injector and more technical expertise, CECT stands out as a cost-effective imaging option, particularly in LMICs. Our findings support the broader use of CECT as a frontline triage modality in acute stroke care and highlight the potential to expand equitable access to life-saving thrombectomy treatment worldwide.