Role of pulmonary perfusion magnetic resonance imaging for the diagnosis of pulmonary hypertension: A review

Miriam Lacharie, Adriana Villa, Xenios Milidonis, Hadeer Hasaneen, Amedeo Chiribiri, Giulia Benedetti

Abstract Among five types of pulmonary hypertension, chronic thromboembolic pulmonary hypertension (CTEPH) is the only curable form, but prompt and accurate diagnosis can be challenging. Computed tomography and nuclear medicine-based techniques are standard imaging modalities to non-invasively diagnose CTEPH,however these are limited by radiation exposure, subjective qualitative bias, and lack of cardiac functional assessment. This review aims to assess the methodology, diagnostic accuracy of pulmonary perfusion imaging in the current literature and discuss its advantages, limitations and future research scope.

Key Words: Pulmonary perfusion MRI; Pulmonary hypertension; Dynamic contrast enhanced magnetic resonance imaging; Chronic thromboembolic pulmonary hypertension;Computed tomography pulmonary angiography; Chronic thromboembolic disease

INTRODUCTION

Pulmonary perfusion is crucial for the assessment of lung function, since proper pulmonary blood flow (PBF) and pulmonary ventilation are prerequisites for effective gas exchange. Pulmonary perfusion can be assessed by measuring mean pulmonary artery pressure (mPAP), pulmonary venous resistance (PVR), and pulmonary artery wedge pressure(PAWP) during right heart catheterization (RHC).

This normal pulmonary perfusion and ventilation pattern can be altered by several physiological and pathological conditions, including vascular (as in cases of chronic pulmonary thromboembolic disease - CTED) and parenchymal alterations (as in cases of lung fibrosis)[1]. These situations can lead to pulmonary hypertension (PH), PH is defined as mPAP of > 20 mmHg and PVR > 2.0 Woods Units[2].

PH is not a single entity, but rather a spectrum of disorders and can be due to multifactorial underlying pathophysiological mechanisms. Current guidelines classify PH in five types based on clinical presentation and aetiology:pulmonary arterial hypertension (PAH), PH due to left heart disease, PH due to lung disease, PH due to CTED (CTEPH)and PH due to unknown mechanisms (e.g., associations include sarcoidosis, sickle cell anaemia, metabolic disorders, and others). Among all PH subtypes, CTEPH is the only treatable form; therefore, early, and accurate differential diagnosis is critical for therapy and prognosis[3,4].

CTEPH is caused by the chronic occlusion of pulmonary arteries due to thromboembolic material with subsequent increased mPAP and PVR, potentially leading to right heart failure and death if left untreated[5,6]. Risk factors for pulmonary thromboembolic vessel occlusion are age over 70, female gender, pulmonary embolism at first venous thromboembolic event, deep vein thrombosis, chronic obstructive pulmonary disease, heart failure and atrial fibrillation[7]. This article will discuss contemporary diagnostic methods to diagnose PH, what methods have been researched so far in medical literature and their diagnostic accuracy. It further explores advantages and limitations of current methods routinely used in clinical practice. It also discusses newly emerging techniques and future horizons of research in PH diagnosis.

Current imaging modalities used for PH diagnosis

Currently, the differential diagnosis of PH involves a multimodality approach including RHC, echocardiography, nuclear medicine based planer ventilation/perfusion scintigraphy scan (V/Q), single photon emission computed tomography(SPECT) scintigraphy and more recently a hybrid approach of combining SPECT and conventional computed tomography (SPECT-CT) scan[8]. All nuclear medicine techniques involve intravenous injection of radionuclide which emits gamma rays after reaching the region of interest in the body. These gamma rays are detected by gamma cameras providing functional assessment of lungs.

Planer scintigraphy has been largely replaced by SPECT because in SPECT gamma cameras rotate over a 360-degree arc around the patient providing three dimensional images which are crucial to detect small segmental perfusion defects.

Recently, hybrid SPECT-CT has also been used for clinical practice. It involves two different types of scans at the same time, a nuclear medicine scan involving radionuclide aerosol of technetium-99m diethylenetriaminepentaacetic acid delivered through a non-breathing mask which reaches to the distal tracheobronchial tree to reflect regional ventilation.Second part of study is an intravenous injection of TC-99m macro aggregates which gets lodged in pulmonary precapillary arterioles to provide pulmonary perfusion assessmentviagamma camera[9,10]. At the same time, conventional CT scanner built in with the SPECT scintigraphy machine provides the anatomical details. Consequently, hybrid SPECT-CT can be considered superior to planar scintigraphy and SPECT alone.

Additionally, computed tomography pulmonary angiography (CTPA) has also been demonstrated to accurately detect pulmonary embolism, however it lacks the ability of dynamic pulmonary perfusion to detect perfusion defects at segmental and sub segmental pulmonary level. CTPA involves intravenous injection of iodine-based contrast injection and imaging of pulmonary arteries for the detection of contrast filling defect denoting pulmonary embolism.

Sensitivity and specificity of CTPA for pulmonary embolism varies from 83% to 100% and with recent advancements in SPECT the potential for routinely used modalities for CTEPH detection has been validated for pulmonary perfusion defect detection[11,12]. This is made possible by the high image resolution offered by multi slice CT scanning, allowing an accurate assessment of the location and severity of thromboembolic material up to the sub-segmental level, providing crucial information for the assessment of patients with CTEPH considered suitable for pulmonary endarterectomy (PEA)[13]. Recently, dual energy computed tomography (CT) has further improved the assessment of patients with PH, thanks to the assessment of regional pulmonary perfusion with iodine perfusion maps[14].

Visual assessment of SPECT scintigraphy, SPECT-CT scintigraphy and CTPA images by expert radiologists is the routine clinical practice to identify partially or completely occluded pulmonary vessels and to assess lung perfusion.

Limitations of CTPA, V/Q scan and SPECT scintigraphy

Routine clinical tests for PH diagnosis are limited by radiation exposure causing carcinogenic stochastic effects (long-term cancer-causing effects after radiation exposure) and rarely deterministic effects (acute high dose exposure if radiation dose exceeds 100 mGy leading to skin erythema, organ damage/failure)[15]. However, these effects are extremely rare due to sophisticated equipment and specialised radiation dose lowering parameters.

The concern is exacerbated for women of child-bearing age, particularly considering the high female-to-male ratio(4.8:1) of incidence and prevalence of PH worldwide and harmful effects on foetal organogenesis[16,17]. Moreover,periodical follow-up to assess the disease progression is vital for PH management, imposing further radiation exposure.

The average effective radiation dose of V/Q scan has been estimated as 1.29 mSv, breast absorbed dose 0.37 mGy,uterus-absorbed dose 0.46 mGy and foetal absorbed doses of 0.40 mGy[18]. SPECT scintigraphy uses radionuclide doses of 25 to 30 MBq for ventilation and 100-120 MBq is usually required for a successful perfusion study[9,19].

Another limitation of current techniques like CTPA is ionising radiation exposure. The effective radiation dose of CTPA has been estimated to be 21 mSv and breast absorbed dose from 44 mGy to as high as 70 mGy to radiosensitive breast tissue[18,20]. Potential carcinogenic ionising radiation exposure and the need for frequent follow-up assessments by conventional tests for PH patients are a few limitations of current imaging tests for pulmonary embolism-based PH diagnosis[21].

Furthermore, as PH can inevitably affect cardiac functions, a single test including the assessment of left and right ventricular function and myocardial tissue characterisation would be desirable for a comprehensive assessment.

In a relevant study, researchers validated that cardiac magnetic resonance imaging (MRI) derived pulmonary artery pressure ejection fraction, RV stroke volume, cardiac output, ventricular mass index and pulmonary blood flow in nontreated lobes correlated with pulmonary blood flow changes in treated lobes post balloon pulmonary angioplastyP<0.05. Furthermore, utility of pulmonary artery pulsatility and pulmonary artery flow measurement by cardiac MRI as an early marker of pulmonary hypertension has also been verified in another prospective study[22]. Therefore, these studies emphasize the significance of a comprehensive, non-invasive and ionising radiation free cardiopulmonary assessment using cardiac magnetic resonance and pulmonary MRI in a single setting.

Finally, in clinical practice these imaging tests are assessed visually only, which is prone to an observer-dependent evaluation, and the availability of quantitative analysis techniques could provide additional value.

Need of improved imaging assessment of CTEPH

Unfortunately, CTEPH retains high mortality, and the mean life expectancy has been reported as less than three years if left untreated[3,23]. Currently, CTEPH can be treated with PEA, which consists of the surgical resection of the thromboembolic material. Before PEA, the three-year mortality of CTEPH was over 50%[24]. It has been demonstrated that PEA can lead to significant improvements in hemodynamic and exercise capacity[23]. Delayed treatment however is still associated with poor prognosis, therefore early detection of CTEPH is critical[25].

A prospective European study of 679 patients with CTEPH, of which 386 underwent PEA, reported an in-hospital mortality rate as low as 4.7% and 1-year mortality of 7%, while the exercise capacity and pulmonary vascular resistance were significantly improved[26]. In another study, the mortality rate in 1500 consecutive CTEPH patients post PEA was investigated between 1999 and 2010. Outcomes were compared between historical cohorts and more recent patients, who received surgical treatment for distal disease affecting smaller pulmonary vessels. A significant reduction of mortality rate from 5.2% to 2.2% was observed, despite the more distal disease, reflecting a continuous improvement in surgical techniques and highlighting the need for more sensitive tests capable of identifying distal disease[27].

Most of CTEPH patients in these studies stated remarkable relief from symptoms after PEA and a return to near normal hemodynamic values[26-28]. These promising results of surgical treatment support the quest for new diagnostic methods capable of early differential diagnosis and further developments of surgical techniques for targeted intervention.Moreover, improved diagnostic tools are necessary for differentiating CTEPH from the other four types of PH because treatment pathways vary accordingly. Currently, this differentiation is based only on imaging tests that employ ionising radiations.

Recently, pulmonary perfusion MRI has been emerged as a powerful tool for the assessment of pulmonary hypertension, and particularly for the differential diagnosis of CTEPH[29]. Perfusion MRI can be performed using dynamic contrast imaging or without contrast using arterial spin labelling[30].

The aim of this review article is to provide an overview of the published studies on lung perfusion MRI, highlighting the limitations of standard tests and the role of perfusion MRI as a diagnostic tool for PH. Moreover, the review focuses on the methodology, advantages, challenges, and future directions for pulmonary perfusion MRI and its potential contribution to overall healthcare policies.

LITERATURE SEARCH

A comprehensive literature search was performed using, PubMed, EMBASE and Medline databases. To develop the literature search question, a PICO framework was adopted: Population; PH patients, Intervention; MRI, Comparison;SPECT/CTPA, Outcome; diagnostic accuracy. Original prospective and retrospective studies were included in the literature search with or without direct comparison of pulmonary perfusion MRI to gold standard tests for the diagnosis of PH[31-55] in Table 1. Tables 2-4 provide the PICO framework, eligibility criteria for included studies and facet analysis with search strategy respectively. Figure 1 demonstrates the PRSMA flow chart for literature search.

Note that the acronym “Dynamic contract enhanced MRI (DCE-MRI)” was frequently used in the selected studies instead of pulmonary perfusion MRI.

Pulmonary perfusion MRI - methodology

MRI uses the interaction between a magnetic field, hydrogen ions in the body and radiofrequency pulses to generate diagnostic images in multiple planes with high spatial resolution and without the need for ionising radiation. An MRI sequence involves repeated application of radiofrequency pulses to the magnetized hydrogen ions in the body and measuring the MRI signals while the hydrogen ions are demagnetized over time (echo time, TE)[56]. Because of the inherent difference in magnetic susceptibility between lung parenchyma and air, MRI of lungs is prone to wider frequency distribution and phase dispersion, leading to suboptimal noisy images[57]. To mitigate this susceptibility effect, recent advancements in lung MRI are ultra-short echo and zero echo time acquisitions, allowing accelerated scan time whilst preserving optimal image quality[58,59].

For fast pulmonary vasculature imaging, gradient pulse sequences with short radiofrequency repetition time and short time to receive the signal are typically employed, which allows acquisition of dynamic images of the lungs in real time as contrast agent flows through the pulmonary vasculature[60] (Figure 2[34]).

A commonly used gradient echo pulse sequence for perfusion MRI is 3-dimensional fast low angle shot (FLASH) MRI acquiring dynamic perfusion images of lungs[61]. The dynamic visualisation of the first pass of the contrast agent across the pulmonary circulation is a crucial requirement for the assessment of lung perfusion[61,62].

Several studies have used 3D FLASH perfusion MRI for the detection of chronic thromboembolic disease[33,34,37,63].However, 3D FLASH perfusion MRI is limited by reduced image resolution and anatomical coverage[64,65]. This is due to the under sampling of the MRI k-space in FLASH MRI imaging. This under sampling of data leads to fast imaging at the expense of image resolution and anatomical coverage.

Ideally, a combination of high temporal resolution (fast imaging) – capable to resolve the dynamic first-pass of the contrast agent, full coverage of the lung fields, and high spatial resolution to capture small pulmonary defects would be required for the optimal differential diagnosis of CTED.

Recently, simultaneous multi-slice (SMS) balanced steady state free precession (b-SSFP) imaging has emerged to solve this problem, allowing better anatomical coverage, increased image resolution and accelerated scan time enabling acquisition of multiple slices simultaneously[65,66]. Therefore, it potentially leads to better visualisation of small pulmonary perfusion defects which can be easily missed in case of sub optimal image quality.

SMS b-SSFP has been recently validated for cardiac perfusion imaging where it doubled spatial coverage by acquiring 6 slices at the same time it takes for conventional sequences to acquire 3 slices, while preserving in-plane resolution[65].The use of SMS b-SSFP for pulmonary perfusion imaging has not been extensively researched, yet further research is warranted considering its promising results for increased anatomical coverage with high image resolution and shorter scan time for cardiac perfusion.

Additionally, pulmonary perfusion MRI involves antero-posterior coronal images covering both lungs and image acquisition on inspiration to minimize respiratory artefacts. To assess perfusion defects, a gadolinium-based MRI contrast injection is used, which offers hyper-intense or hypo-intense signals depending on the amount of blood perfusion in the region of interest over time (dynamic imaging). Recently, free breathing lung MRI was also validated for image quality and reproducibility, showing promise for increased patient comfort during scanning of PH and patients with other lung diseases causing poor breath holding[67,68].

Pulmonary perfusion MRI images can be assessed qualitatively and quantitatively.

Qualitative (visual) assessment of pulmonary perfusion MRI

Current clinical practice is to visually assess the lung perfusion MRI images during the passage of contrast agent through the pulmonary circulation. Dynamic pulmonary perfusion MRI leads to the visualisation of pulmonary arteries’ contrast distribution: a pulmonary segment with normal perfusion shows a good and homogeneous perfusion, while the lung territory supplied by a pulmonary artery with partial or complete occlusion due to thromboembolic material will show reduced and delayed perfusion. Impaired lung perfusion can be visually assessed as a wedge-shaped perfusion defect on the pulmonary perfusion images.

However, this observer based qualitative assessment is prone to inter- and intra-observer variations leading to subjective interpretations[69,70].

Quantitative assessment of pulmonary perfusion MRI

To mitigate subjective bias, pulmonary perfusion MRI also allows quantification of perfusion parameters: Pulmonary blood flow, pulmonary blood volume and mean transient time of blood flow (Figures 3 and 4[33,50]) which requires tracing of the contrast agent passing through the tissue of interest and an arterial input function (AIF) curve corresponding to the signal intensity in the blood pool over time[66].

Perfusion quantification is based on the principles of the indicator dilution theory, relating the AIF and tissue enhancement curvesviathe process of deconvolution[71-73], and is often used for the measurement of myocardial blood flow[74]. Pulmonary blood volume (PBV) and mean transit time (MTT) can be subsequently estimated according to the central volume theorem[33,73]. Quantification is, however, significantly impacted by the nonlinear relationship of themeasured signal with contrast agent concentration leading to errors in quantitative metrics.

Table 1 Characteristics of included studies

thromboembolism(APTE)gradient sequence,MPAP, PVR measured by RHC. RV/LV diameter ratio, APTE index measured by CT& MRA P < 0.05), APTE indexes from all modalities proved significant predictors for differentiating APTE patients Stein et al[41],2010 Prospective,qualitative 371 371 adults with diagnosed or excluded pulmonary embolism- (PIOPED III)To assess performance of MRA and venography for pulmonary embolism detection MRA was compared with CTPA, V/Q scan,venous ultrasonography, D-dimer assay,and clinical assessment,Qualitative assessment by expert reader only Chi-square test ANOVA Technically adequate images for MRA: SE:78%, SP: 99%Kang et al[42],2011 Prospective,qualitative 35 35 PAH patients(Mean age 44 yr)To assess if Cardiac MRI based pulmonary artery distensibility index correlates with RHC estimates for PAH Pulmonary artery distensibility indices were derived from transverse view MRI and compared with PVR using RHC Correlation Non-invasive MRI based pulmonary artery distensibility index correlates with RHC based estimates P < 0.001 Ohno et al[43],2012 Prospective,qualitative 24 Response group=13,Non-response group=11, 12 females & 12 males mean age 68 yr ± 8.6 CTPA, MRA & DCEMRI comparison for treatment response in inoperable CTEPH patients PBF, PBV, MTT measured by DCE-MRI using 3-dimensional spoiled gradient sequence, RV/LV diameter ratio and embolic burden measured by CTPA &MRA Mean of student T test,Correlation,ROC curve analysis,McNemar’s test DCE-MRI SP = 90%,AC = 95%, CTPA SP= 36%, AC = 70%,MRA SP = 54%, AC =79%Ley et al[44],2013 Prospective,qualitative 20 PAH or CTEPH patients Controls 10,Training group 10 To evaluate if training improves pulmonary perfusion in PH as assessed by MR perfusion imaging Training group received in hospital exercise training while control group received conventional rehabilitation. 6 min walk test,PBF, PBV, MTT & peak flow velocity measured by MR perfusion were assessed for both groups from baseline to 3 wk Mann-Whitney-Wilcoxon test,Spearman correlation coefficient Training group had significantly improved 6-min walk test, MR flow and MR perfusion Rajaram et al[45], 2013 Prospective,qualitative 132 78 CTEPH patients To compare the diagnostic accuracy of perfusion MRI for CTEPH Vs. CTPA and V/Q Pulmonary perfusion MRI using time resolved 3-Dimensional spoiled gradient and pulmonary MRA were compared with CTPA and V/Q Not given SE, SP in %, MRI: 97,92, V/Q: 96, 90,CTPA: 94, 98 Revel et al[46],2013 Prospective,qualitative 274 274 suspected PE patients To evaluate unenhanced,enhanced perfusion and MR angiography for PE detection Unenhanced steady state free precession or SSFP, fast spoiled gradient echo for perfusion MRI and MR angiography were compared with CTPA Chi-squared Kappa statistics Kappa agreement MRA = 0.77,Perfusion MRI = 0.51,Unenhanced MRI =0.62 Sugimoto et al[47], 2013 Prospective,qualitative 34 34 congenital heart disease patients To assess if velocity encoded cine imaging can measure pulmonary artery pressure in children with congenital heart disease Pulmonary blood flow(QP), systemic blood flow (QS), acceleration time, ejection time,peak velocity, and maximal change in flow rate during ejection (MCFR) were measured by velocity encoded MRI and RHC Velocity encoded MRI correlated strongly with RHC for QS,right to left QP ratio and QP/QS.Suggesting usefulness of MRI for pulmonary artery pressure measurement Schoenfeld et al[48], 2015 Prospective,qualitative 64 64 ruled out or confirmed PE patients To compare perfusion weighted Fourier decomposition or PW-FD to DCE-MRI for PE detection Time resolved angiography with stochastic trajectories or TWIST for DCE-MRI was used and compared with PW-FD Qualitative only, Kappa statistics For PW-FD per patient basis, SE =100%, SP = 95%, PPV= 98%, NPV = 98%,Intraobserver k =0.96, Interobserver k= 0.96 Ingrisch et al[49], 2016 Prospective,qualitative 18 8 acute PE, 10 controls DCE-MRI evaluation for acute PE Qualitative assessment of presence and Cohen’s kappa, Fisher’s SE: 87-93%, SP: 90-95%, PPV: 87-93%,

AC: Diagnostic accuracy; APTE: Acute pulmonary thromboembolism; BPA: Balloon pulmonary angioplasty; CI: Confidence interval; COPD: Chronic obstructive pulmonary disease; CTEPH: Chronic thromboembolic pulmonary hypertension; CTPA: computed tomography pulmonary angiography; CO:Cardiac output; CV: Coefficient of variation; DCE-MRI: Dynamic contrast enhanced magnetic resonance imaging; %DL(CO): Diffusing capacity of carbon monoxide; LV: Left ventricle; LCI: %(lung clearance index), MRA: Magnetic resonance angiography; MTT: Mean transit time; NPV: Negative predictive value; PAP: pulmonary artery pressure; PBF: Pulmonary blood flow; PBV: Pulmonary blood volume; PE: Pulmonary embolism; FVC: % Percent predicted forced vital capacity; PPV: positive predictive value; PVR: Pulmonary venous resistance; ROC: Receiver operating curve; SE: Sensitivity; SP: Specificity;sPAP: Systolic pulmonary artery pressure; SPECT: Single photon emission Computed tomography; SPSS: Statistical package for social sciences; RV: Right ventricle; V/Q: Ventilation perfusion scan.

Non-linearity is caused by spatial signal variations due to variations in sensitivity profiles of MRI surface coils, contrast bolus related T2* decay and subsequent signal loss[66]. To mitigate this effect, a dual-contrast bolus approach can be used, where a dilute (low concentration) bolus is injected for AIF measurement and a neat (high concentration) bolus is subsequently injected for signal intensity measurement in the lung tissue[75]. Researchers reported that the dual-bolus approach is feasible for pulmonary perfusion MRI, reducing the inherent nonlinearity between contrast concentration and signal intensity[76].

Recently, the dual-sequence approach was introduced to avoid the complexity of injector set up and issues of nonlinearity for cardiac perfusion MRI[77]. The dual-sequence approach is more practical in busy clinical settings, as it involves injection of a single contrast bolus. The sequence acquires a low resolution slice by FLASH readout for sampling the AIF, followed by acquisition of high resolution short-axis slices using SSFP or FLASH readout[78]. This dual-sequence technique has been validated for cardiac perfusion imaging for its linearity for signal conversion to contrast bolus concentration[66].

Clinically, MRI perfusion quantification permits user-independence quantification of PBF, PBV and MTT at the global,lobar, and segmental pulmonary levels, minimising inter- and intra-observer qualitative bias[69]. Quantification of perfusion parameters allow narrowing down the differential diagnosis of lesions associated with increased or decreased PBF or PBV, (as in the case of chronic thromboembolic partial or complete obstruction of pulmonary vasculature[79].

Diagnostic accuracy of pulmonary perfusion in the literature

Evaluation of pulmonary thromboembolic disease by lung perfusion MRI is a relatively new imaging technique, and its diagnostic potential has been reported comparable with the routinely performed tests of CTPA and SPECT scintigraphy[43,50,63]. The diagnostic accuracy of pulmonary perfusion has been tested qualitatively and quantitatively.

Qualitative assessment: An early feasibility study looking at the diagnostic accuracy of perfusion MRI for detecting lung perfusion defects in patients with pulmonary embolism revealed a good inter-modality agreement with V/Q scan by correctly identifying 16 out of 18 perfusion defects[31].

The diagnostic accuracy of pulmonary perfusion MRI has been demonstrated comparable to V/Q scan measured by the Kappa statistic (k) for inter-observer and inter-modality agreement (inter-observer agreement: k = 0.63 and intermodality agreement: k = 0.79) in another prospective qualitative study assessing the treatment response of inoperable CTEPH patients[34].

A prospective qualitative study on 78 CTEPH patients assessing the diagnostic accuracy of pulmonary perfusion MRI concluded a sensitivity of 97% and specificity of 92% compared to V/Q and CTPA[45]. Another qualitative comparison for evaluating the diagnostic potential of pulmonary perfusion MRI for the detection of pulmonary embolism reported a sensitivity of 87%-93% and a specificity of 90%-95% with an inter-observer agreement of k = 0.77[49]. Similarly, another study prospectively comparing the diagnostic potential of perfusion MRI, perfusion SPECT and CTPA revealed sensitivity 100% and specificity 81% with inter-observer agreement of k=0.88 for perfusion MRI compared to interobserver agreement of k = 0.80 (P< 0.0001) for SPECT scintigraphy[50].

Quantitative assessment: Prospective quantitative studies affirm the diagnostic potential of pulmonary perfusion MRI by measuring the perfusion parameters of PBF, PBV and MTT for normal, CTEPH and PH of other aetiologies leading to early differential diagnosis and better disease management[37,39,40,43].

Pulmonary perfusion parameters of PBF, PBV and MTT for healthy volunteers and PH patients were quantitatively assessed in a prospective study. Pulmonary perfusion parameters were significantly lower for PH patients accurately differentiating the two study groups[33]. Furthermore, pulmonary perfusion MRI diagnostic accuracy was confirmed in another relevant study where significant results were reported for all three pulmonary perfusion MRI parameters correctly differentiating the two study groups and aiding the differential diagnosis of primary pulmonary hypertension[37].

Additionally, the capability of pulmonary perfusion MRI for the evaluation of disease severity and progression of PAH was compared with pulmonary function test and Doppler echocardiography. Pulmonary perfusion MRI was positively correlated with diffusing capacity of carbon monoxide %DL(CO) measured by pulmonary function test. In the same study, pulmonary perfusion MRI parameters of PBF, PBV and MTT were also positively correlated with mean pulmonary artery pressure and systolic pulmonary artery pressure measured by Doppler echocardiography[39]. This significant correlation implies promising diagnostic potential of pulmonary perfusion MRI for detecting PAH when compared to routine tests.

In another quantitative prospective study, specificity and accuracy of pulmonary perfusion MRI were compared with MDCT and magnetic resonance angiography (MRA) for PH diagnosis. Specificity and accuracy were significantly higher for pulmonary perfusion MRI when right ventricle/Left ventricle diameter ratio and acute pulmonary thromboembolic indices were compared for perfusion MRI, MDCT and MRA (P< 0.05). This implies comparable diagnostic capability for pulmonary perfusion MRI to routine tests for PH diagnosis. In this study, PBF was reported as a more accurate perfusion MRI parameter than PBV and MTT[40].

Table 3 Inclusion and exclusion criteria for study selection

Table 4 Facet analysis and search strategy

Moreover, researchers compared the accuracy of treatment response in inoperable CTEPH patients using pulmonary perfusion MRI, CTPA and MRA. This comparative prospective study found higher diagnostic accuracy (95%) and specificity (90%) for pulmonary perfusion MRI compared to CTPA to differentiate treatment responders and nonresponders[43]. A recent quantitative study assessed the regional PBF pre and post PEA using pulmonary perfusion MRI and compared it with exercise capacity post PEA for CTEPH patient. They concluded a positive correlation between perfusion MRI for identifying improvement in PBF in the lower lungs and exercise capacity after successful surgery[80].

Figure 1 PRISMA flow chart for literature search.

This literature search concludes the approving potential of pulmonary perfusion MRI for the differential diagnosis of PH. However, most of these comparative studies were single centre with small sample size, therefore further large-scale multicentre research is recommended.

Hyperpolarised inhaled contrast agent pulmonary MRI

Hyperpolarised gases like3He,129Xe and Oxygen-enhanced1H have also gained popularity for assessing pulmonary ventilation and perfusion in lung diseases. A retrospective study on 15 patients was conducted to assess regional lung ventilation as part of pre-operative planning of lobectomy, pneumonectomy or lung volume reduction surgery. This study reported good agreement between3He MRI and scintigraphy ventilation imaging to diagnose ventilation defects[81]. Another case study of CTEPH patient demonstrated the use of3He ventilation scan combined with dynamic contrast enhanced perfusion MRI to assess the treatment response of pre and post PEA[82].

129Xe is of particular interest for CTEPH diagnosis because129Xe has an ability to dissolve in pulmonary tissues. After inhalation,129Xe is easily soluble in lung tissues and blood leading to frequency shift which can be quantitatively measured by MR spectroscopy[83].129Xe has been tested to demonstrate a reduced uptake of red blood cells in CTEPH and pulmonary vascular disease patients indicating perfusion defects[84,85]. Another study on 10 PAH patients,129Xe MRI showed reduced RBC amplitude oscillations on MRI spectroscopy implying its potential to diagnose pulmonary vascular diseases from other lung disorders[86]. These initial results reflect the potential of combining hyperpolarised gas MRI imaging with cardiopulmonary MRI to comprehensively assess the structural and functional ventilation and perfusion pulmonary disorders.

However, hyperpolarised inhaled contrast agent pulmonary MRI is still an emerging modality and further large-scale research is needed to strengthen its diagnostic potential to conventional perfusion scintigraphy.

Advantages of pulmonary perfusion MRI

Ionising radiation free diagnostic test: Screening and follow-up scans to monitor the disease progression is a crucial part of management strategy for patients with PH. In a prospective longitudinal study in a cohort of 675 emergency department patients researchers concluded that at least one third of patients undergoing CTPA were called for a follow up scan within 5 years, with one fifth of patients being women younger than 40 years old[21]. In addition, it is estimated that 5% to 11% of CTPA scans for the diagnosis of CTEPH are repeated due to artefacts and technical reasons, which further exacerbate the risk of excessive radiation dose in another study[87]. This unnecessary ionising radiation exposure can be reduced by using pulmonary perfusion MRI at screening and diagnostic stages of CTEPH leading to better patient healthcare strategy.

An all-inclusive test: Besides lung perfusion imaging, the cardiac assessment is also important in patients with PH.Cardiac involvement in PH is an inevitable step in the disease progression and PH due to left heart disease is the most common type[88,89]. It is reported that more than half of heart failure patients suffer from PH[90]. The pressure increase in the pulmonary arteries leads to an increase in the RV pressure and cardiac remodelling, correlated with increased mortality. RV end-systolic volume index has been reported as an independent predictor of PAH prognosis in a study using 288 derivation cohort and 288 validation PAH cohort[91].

Figure 2 Comparing digital subtraction angiography, pulmonary perfusion magnetic resonance imaging and magnetic resonance angiography for identifying perfusion defects in chronic thromboembolic pulmonary hypertension patient. A: Digital subtraction angiography showing perfusion defects (arrows) in left upper lobe and right lower lobe due to thromboembolic occlusion; B: Magnetic resonance imaging perfusion image at peak enhancement showing hypo-intense perfusion defects at same locations; C: Magnetic resonance angiogram of the same patient demonstrating dark thromboembolic material in the right pulmonary artery corresponding to perfusion defects at the same location. Citation: Nikolaou K, Schoenberg SO, Attenberger U, Scheidler J,Dietrich O, Kuehn B, Rosa F, Huber A, Leuchte H, Baumgartner R, Behr J, Reiser MF. Pulmonary arterial hypertension: diagnosis with fast perfusion MR imaging and high-spatial-resolution MR angiography--preliminary experience. Radiology 2005; 236: 694-703 [PMID: 15994997 DOI: 10.1148/radiol.2361040502].

Figure 3 Comparing single photon emission computed tomography and dynamic contract enhanced perfusion magnetic resonance imaging for identifying perfusion defects in chronic thromboembolic pulmonary hypertension patient. Left to right, single photon emission computed tomography and Perfusion magnetic resonance imaging (MRI) at peak enhancement of a chronic thromboembolic pulmonary hypertension patient showing perfusion defects in right and left lower lobes. On the right, semi-quantitative pulmonary blood flow, pulmonary blood volume & mean transit time maps acquired from perfusion MRI demonstrating reduced flow at the same locations. Note left to right scale depicting lower to higher pulmonary blood flow values. Citation: Johns CS,Swift AJ, Rajaram S, Hughes PJC, Capener DJ, Kiely DG, Wild JM. Lung perfusion: MRI vs. SPECT for screening in suspected chronic thromboembolic pulmonary hypertension. J Magn Reson Imaging 2017; 46: 1693-1697 [PMID: 28376242 DOI: 10.1002/jmri.25714].

Figure 4 Comparing computed tomography chest and dynamic contract enhanced perfusion magnetic resonance imaging for identifying perfusion defects in pulmonary hypertension patient with COPD. A: Chest computed tomography showing lobular emphysema in the lungs; B-D:Quantitative perfusion maps (anterior to posterior) showing heterogeneously reduced pulmonary blood flow, pulmonary blood volume and mean transit time respectively representing thromboembolic occlusion in respective pulmonary segments. Note on the right-hand side, quantitative scale from top to bottom showing higher to lower quantitative values. Citation: Ohno Y, Hatabu H, Murase K, Higashino T, Kawamitsu H, Watanabe H, Takenaka D, Fujii M, Sugimura K. Quantitative assessment of regional pulmonary perfusion in the entire lung using three-dimensional ultrafast dynamic contrast-enhanced magnetic resonance imaging: Preliminary experience in 40 subjects. J Magn Reson Imaging 2004; 20: 353-365 [PMID: 15332240 DOI: 10.1002/jmri.20137].

Moreover, another relevant study validated that cardiac MRI based pulmonary artery distensibility index correlates positively with the right heart catheterization estimates for PAH, suggesting non-invasive cardiac MRI a valuable tool for PAH diagnosis[42].

Additionally, cardiac magnetic resonance imaging has also been used to assess pulmonary artery pressure measurements in children with congenital heart disease. Researchers found velocity encoded flow cardiac MRI images strongly correlated with right heart catheterization for systemic blood flow (QS), right to left pulmonary blood flow (QP)ratio and QP/QS ratio[47]. This suggests usefulness of cardiac MRI for pulmonary artery pressure measurement by using the standard SSFP cine imaging and phase contrast flow imaging.

Pulmonary perfusion can be combined with cardiac MRI, including cardiac cine images for structural and functional assessment of heart chambers, blood flow assessment by phase contrast imaging, MRA, post contrast late gadolinium enhancement imaging for myocardial fibrosis/scarring and four-dimensional phase contrast pulmonary artery flow scan.A comprehensive cardio-pulmonary assessment can be achieved in a single examination during the same visit, leading to better use of health resources and limiting the patients’ stress[83].

Quantification mitigates observer bias: MRI pulmonary perfusion offers quantification of PBF, PBV and MTT with improved diagnostic accuracy. In a study assessing the diagnostic accuracy of qualitative and quantitative cardiac MR perfusion in coronary artery disease, the level of training was reported as a major determinant for the accuracy of qualitative assessment (the diagnostic accuracy of level-3 operators was 83.6%, level-2 operators 65.7% and level-1 operators 55.7%,P< 0.001), while the diagnostic accuracy of quantitative assessment was 86.3% (significantly higher than level-2 and level-1 operators), highlighting the usefulness of operator-independent quantification techniques[70].

Moreover, higher consistency and reproducibility results were achieved using quantification techniques, and an increased sensitivity (from 77% to 83%) was reported when using quantitative measures for the evaluation of myocardial perfusion[92]. Researchers also studied MRI signal abnormalities in the brain and reported higher reproducibility and accuracy of quantification techniques compared to qualitative analysis by expert radiologists[93]. The quantification of MRI perfusion parameters mitigates this observer bias, which is not achievable with current tests to diagnose PH.

Contrast nephrotoxicity comparison

CT scan uses iodine-containing contrast media which can unfortunately cause serious complications, such as contrast induced nephropathy. Gadolinium-based contrast agents are considered safer than iodinated contrast media, thus they can be safely used also in patients with impaired kidney function[94]. CT contrast also involves a huge dose and contrast related discomfort has also been reported. Additionally, the iodine-based contrast can potentially worsen the thyroid function, on the other hand gadolinium based MRI contrast agents have been reported safer for pregnant patients and patients with compromised renal functions[95]. Some other contraindications for CT contrast injection include multiple myeloma and metformin.

CONSIDERATIONS REGARDING PULMONARY PERFUSION MRI

Claustrophobia and magnet safety related issues

An MRI scanner is a long, tunnel-like enclosed chamber where patients need to lie down still for a long time (minimum 40 min). Patients are further fastened by the radiofrequency coil placed on the chest. The MRI scanner also produces loud banging noise (110 decibels at its loudest)[96]. Lying down still for a long time in such an enclosed chamber and tolerating the loud noise can be challenging for claustrophobic patients or patients with hearing problems. Magnet related safety for medical devices’ including cardiac pacemakers, aneurysmal clips, intra orbital metallic injury and some other incompatible metallic implants are some other limitations of MRI scanning.

Consideration about image resolution, scan time and respiratory artefacts

In comparison to computed tomography, MRI is prone to breathing artefacts and it inherently retains low image resolution for moving organs like heart and lungs. Scan time is another limitation for MRI, longer scan time leads to patient discomfort. To mitigate, advanced motion correction techniques of parallel imaging, patient training for breath hold, respiratory/ electrocardiogram gating for cardiac MRI and saturation band to supress signal from moving chest wall and heart motion have been adapted[97].

Moreover, recently researchers have successfully validated free breathing, real time MRI imaging with short scan time and high image resolution for cardiac perfusion imaging[98,99]. More research is warranted to apply this technique for pulmonary perfusion

Technical challenges of pulmonary perfusion quantification

Pulmonary perfusion quantification is technically challenging. It needs a skilled operator and availability of a specific software package for quantification. Furthermore, it is prone to errors when images have respiratory and motion artefacts. Lack of integration with other clinical findings while quantification is another limitation.

Future directions

The Fleischner Society’s position paper recommends MRI for clinical use for cystic fibrosis, lung cancer staging, lung nodule characterisation and pulmonary hypertension. The position paper endorses further investigating the role of perfusion MRI for the diagnosis of chronic obstructive pulmonary disease and pulmonary parenchymal abnormalities in future[57]. However, MRI imaging for lungs is still limited in clinical and research fields implying the need for further clinically implementable research in this field[57,100]. An ongoing prospective, multicentre, comparative phase III clinical trial (Change-MRI) is currently being carried out to compare the diagnostic accuracy of perfusion MRI and SPECT scintigraphy for CTEPH diagnosis[101]. Pulmonary perfusion MRI has clinical and technical aspects to be considered for future research on this topic.

Clinically, combining cardiac magnetic resonance with pulmonary perfusion imaging can help differentiate lung disorders of multiple aetiologies. 4-dimentional flow MRI imaging to assess pulmonary artery pressure has been demonstrated to aid CTEPH diagnosis[53,102,103]. Moreover, as SMS b-SSFP MRI sequence has been validated for better anatomical coverage and improved resolution for cardiac imaging, its use can be expanded for pulmonary perfusion imaging and specifically thromboembolic vascular lung disorders. Additionally, sophisticated software to quantitatively analyse perfusion parameters can further improve diagnostic accuracy in future[57].

Quantitative analysis of perfusion MRI parameters of PBF, PBV and MTT can be performed manually or automatically.Manual quantification can be time consuming and laborious. To improve efficiency, researchers established the validity of automated quantitative analysis for myocardial perfusion MRI and found it more efficient and reproducible than manual analysis[104-106].

Similarly, the automatic analysis of lung MRI has also been researched using artificial intelligence (AI) and deep learning algorithms for automatic lung segmentation of 20 cystic fibrosis patients. This study found automatic lung segmentation comparable with manual segmentation and capable of accurate estimation of lung volume[107].Researchers also used AI techniques to automatically quantify the neonatal lung structure and compared it with manual segmentation by expert operators. This study found that AI-supported automatic segmentation of lung MRI was accurate and comparable to expert-level accuracy[108]. Therefore, for efficiency of time and consistency of diagnostic parameters,further research on the accuracy of automatic lung segmentation and quantification of perfusion parameters is recommended.

Currently, the dual-bolus approach is preferred in routine clinical practice for perfusion MRI imaging, but the dualsequence approach to account for signal intensity nonlinearity and operator-friendly set up of injector pump in busy clinical settings is also recommended for future research on this topic[66].

CONCLUSION

PH is multifactorial and requires timely multimodality differential diagnosis for better disease prognosis. Pulmonary perfusion MRI can be considered a reliable and safe diagnostic imaging tool with comparable diagnostic accuracy to reference gold standards of SPECT, CTPA and V/Q imaging. It offers cardiopulmonary assessment when combined with CMR at the same visit without any radiation exposure making it a comprehensive test for screening, differential diagnosis and follow up scans for PH management. Furthermore, quantification of pulmonary perfusion parameters could improve the reliability and accuracy of diagnostic testing and help mitigate subjective interpretation of imaging findings which is not possible by qualitative assessment alone offered by standard tests.

FOOTNOTES

Author contributions:Villa ADM was responsible for the review concept; Lacharie M drafted the manuscript; Lacharie M provided critical revision of the manuscript for important intellectual content; and all authors critically reviewed the content.

Conflict-of-interest statement:Authors declare no conflict of interest for this article.

Open-Access:This article is an open-access article that was selected by an in-house editor and fully peer-reviewed by external reviewers.It is distributed in accordance with the Creative Commons Attribution NonCommercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: https://creativecommons.org/Licenses/by-nc/4.0/

Country/Territory of origin:United Kingdom

ORCID number:Miriam Lacharie 0000-0002-7282-6894; Adriana Villa 0000-0002-9355-0372; Xenios Milidonis 0000-0002-5446-9459; Amedeo Chiribiri 0000-0003-3394-4289.

Corresponding Author's Membership in Professional Societies:Health and Care Professionalcouncil, RA75559.

S-Editor:Liu JH

L-Editor:A

P-Editor:Zhao S