AJTCCM VOL. 31 NO. 1 2025 19
ORIGINAL ARTICLES: RESEARCH
Background. Pulmonary hypertension (PH) aer tuberculosis is increasingly recognised as important in high-burden tuberculosis settings.
However, the ability of computed tomography (CT) imaging to accurately detect PH remains unclear.
Objectives. To evaluate the performance of standard CT measurements in detecting PH in patients with post-tuberculosis lung disease
(PTLD), and to determine the potential role of CT imaging as a screening tool in this population.
Methods. A retrospective study of patients with PTLD was conducted from January 2019 to September 2021. Adult patients with both a
CT chest scan and an echocardiogram performed within 9 months of each other were enrolled. A diagnosis of PH by echocardiography
was made if the right ventricular systolic pressure (RVSP) was ≥36 mmHg or the peak tricuspid regurgitant jet velocity (TRVmax) >2.8 m/s.
Radiological criteria for PH included a pulmonary artery/ascending aorta (PA/AA) diameter ratio >1, pulmonary artery diameter (PAD)
≥29 mm (males) or ≥27 mm (females), and right ventricle/le ventricle (RV/LV) diameter ratio ≥1.28. Spirometry was also performed.
Results. Of 173 patients with PTLD, 52 met the inclusion criteria. Signicant correlations were found between the CT-measured
PA/AA ratio and RVSP (p=0.0083) and TRVmax (p=0.0582), but not between the CT-measured RV/LV ratio and RVSP (p=0.1729) or
TRVmax (p=0.0749). PAD was also signicantly correlated with RVSP (p=0.0011) and TRVmax (p=0.0023). e PA/AA ratio identied
patients with PH on echocardiography with ~100% sensitivity, 65% specicity and a positive predictive value of 39.1%, indicating a high
potential for false-positive diagnosis. e forced vital capacity was 13.7% lower in patients with PH than in those without (p=0.044); however,
the forced expiratory volume in 1 second was not statistically dierent.
Conclusion. A low PA/AA ratio can be used to rule out the diagnosis of PH in PTLD, but a high PA/AA ratio requires further investigation
for PH.
Keywords. Post-tuberculosis lung disease, pulmonary hypertension, computed tomography scan, echocardiography, pulmonary artery/
ascending aorta ratio.
Afr J Thoracic Crit Care Med 2025;31(1):e1948. https://doi.org/10.7196/AJTCCM.2025.v31i1.1948
Computed tomography chest imaging for the detection of
pulmonary hypertension in patients with post-tuberculosis
lungdisease
M Almubarek,1 MB ChB, MMed (Int Med), FCP (SA) ; E H Louw,1 MB ChB, FCP (SA), Cert Pulmonology (SA) ;
S Grith-Richards,2 MB ChB, MMed (Rad D), FC Rad Diag (SA) ; CAckermann,2 MB ChB, MMed (Rad D), PhD ;
N Baines,1 BSc ; H omson,3 MB ChB, BSc, DTM&H, MRCP ; A J K Pecoraro,4 MB ChB, FCP (SA), Cert Cardiology (SA), PhD ;
C F N Koegelenberg,1 FCP (SA), FRCP, Cert Pulmonology (SA), PhD ; E M Irusen,1 MB ChB, FCP (SA), PhD ;
B W Allwood,1 MB ChB, FCP (SA), Cert Pulmonology (SA), PhD
1 Division of Pulmonology, Department of Medicine, Faculty of Medicine and Health Sciences, Stellenbosch University and Tygerberg Hospital, Cape Town,
SouthAfrica
2 Department of Radiology,Faculty of Medicine and Health Sciences, Stellenbosch University and Tygerberg Hospital, Cape Town, South Africa
3 Cleveland Clinic, London, UK
4 Division of Cardiology, Department of Medicine, Faculty of Medicine and Health Sciences, Stellenbosch University and Tygerberg Hospital, Cape Town,
SouthAfrica
Corresponding author: M Almubarek (marwanalmobark89@gmail.com)
Study synopsis
What the study adds. is study investigated the use of computed tomography (CT) chest imaging to detect pulmonary hypertension
(PH) in patients with post-tuberculosis lung disease (PTLD). It revealed signicant correlations between the CT-measured pulmonary
artery/ascending aorta (PA/AA) diameter ratio and pulmonary artery diameter (PAD), and echocardiographic measures of PH. Notably,
a low PA/AA ratio eectively rules out PH, while a high ratio warrants further investigation.
Implications of the ndings. ese ndings suggest that CT imaging, particularly PA/AA ratio measurements, could serve as a valuable
initial screening tool for ruling out PH in patients with PTLD, particularly in settings with limited access to echocardiography. However, a
high PA/AA in PTLD requires conrmation of PH by other means, owing to a low positive predictive value.
20 AJTCCM VOL. 31 NO. 1 2025
ORIGINAL ARTICLES: RESEARCH
Pulmonary hypertension (PH) is a group of diseases characterised
by elevated pulmonary artery pressure, oen leading to right heart
failure and early mortality.[1,2] e denition of PH was revised in
2018, and a lowered threshold of mean pulmonary arterial pressure
≥20 mmHg was adopted.[3,4] Among the dierent classications of
PH, group 3 PH, secondary to chronic lung disease, is the second
most common subtype.[2,3]
Chronic lung diseases such as chronic obstructive pulmonary
disease (COPD), interstitial lung disease, and combined pulmonary
brosis and emphysema are frequently associated with PH, which
portends a poorer prognosis.[2,3]
Tuberculosis (TB), globally the leading cause of death from a
single infectious agent,[5] primarily aects the lungs. If le untreated,
TB can result in long-term damage, including extensive pulmonary
destruction, chronic airflow obstruction, and destruction of the
pulmonary vascular bed.[6] However, the association between post-TB
lung disease (PTLD) and PH is under-reported in the literature.[1,2,6,7]
The development of PH following TB is believed to involve
mechanisms such as pulmonary vascular remodelling, parenchymal
pathology leading to destruction of the vascular bed, vasculitis,
thrombosis of the pulmonary artery, endarteritis obliterans, and
brosing mediastinitis.[6,8]
Diagnosing PH can be challenging owing to its nonspecic clinical
manifestations.[9] However, timely and accurate diagnosis of PH is
essential for prognosis and treatment planning.[9] Echocardiography,
a non-invasive and widely available imaging modality, plays a
key role in screening for and diagnosis of PH.[10,11] Raised peak
tricuspid regurgitant jet velocity (TRVmax) together with other
echocardiographic ndings has been proposed as a reliable indicator
of the probability of PH, as endorsed by the European Society of
Cardiology (ESC), the European Respiratory Society (ERS) and the
6th World Symposium on Pulmonary Hypertension.[2]
Although echocardiography is the initial screening test of choice,[10,11]
it may present challenges in patients with lung disease, owing to lung
hyperination and diculty in obtaining clear acoustic windows.[12] In
contrast, computed tomography (CT) imaging enables comprehensive
evaluation of the heart, pulmonary arteries, parenchyma and
mediastinum, making it a valuable tool in the evaluation of suspected
PH.[3,4,13] e presence of PH can be identied through anatomical
vascular changes, such as a pulmonary artery/ascending aorta
(PA/AA) diameter ratio >1, pulmonary artery diameter (PAD)
≥29mm in males and ≥27 mm in females,[3,4,13] and right ventricle/
le ventricle (RV/LV) diameter ratio ≥1.28.[3]
CT measurements predict the presence of PH with 70% sensitivity
and 92% specicity compared with right heart catheterisation (RHC).[14]
Echocardiography predicts the presence of PH with a sensitivity of
87% and a specicity of 79% compared with RHC.[15]
While the PA/AA ratio has shown promise in predicting PH in
COPD,[12] its applicability in PTLD remains unclear.[12] e aim of
this study was to investigate whether CT features correlate robustly
with estimates of pulmonary artery pressures obtained through
echocardiography in patients with PTLD.
Methods
Study design and patient selection
In this retrospective observational study conducted between January
2019 and September 2021, patients previously treated for pulmonary
TB were eligible for inclusion if they were >14 years of age and had
had an echocardiogram and a CT scan performed within 9 months
of each other. is time frame was chosen because many patients
have CT chest imaging conducted at secondary hospitals without
simultaneous echocardiograms, frequently with subsequent prolonged
delays in referral to our institution. e study was undertaken in the
Division of Pulmonology at Tygerberg Hospital, a tertiary hospital
in Cape Town, South Africa. Potential participants were identied
from the departmental PTLD registry or from patients attending
follow-up visits. ey were excluded if they had an alternative known
cause for PH (e.g. COPD, interstitial lung disease, obstructive sleep
apnoea, autoimmune diseases, connective tissue disorders, collagen
vascular disorders). In addition, potential participants were excluded
if they had missing data, poor views or unreliable echocardiographic
measurement and assessment, or a scan time interval >9 months, or
if their CT scan was deemed uninterpretable for technical reasons
(e.g.anatomical distortion).
Each participant had spirometry performed, and echocardiographic
and CT chest data were used to assess for features indicative of PH.
e data were interpreted by a pulmonologist, a radiologist and a
cardiologist. Radiological criteria for diagnosis of PH were a PA/
AA ratio >1, PAD ≥29 mm in males and ≥27 mm in females, and an
RV/LV ratio ≥1.28. Echocardiographic criteria for PH were dened
as a right ventricular systolic pressure (RVSP) ≥36 mmHg and a
TRVmax>2.8 m/s. Although the gold standard for the diagnosis of PH
is RHC, it was not required for inclusion. Ethical approval to conduct
this study was obtained from the Human Research Ethics Committee
of Stellenbosch University (ref. no. S22/08/141).
Measuring the PA/AA diameter ratio, PAD, and the
RV/LV diameter ratio
Two radiology specialists independently measured the diameters of
the PA, AA, RV and LV, and the average of the two measurements was
calculated for each diameter and ratio. e diameters of the PA and
AA were measured at a level proximal to the bifurcation of the PA, and
for measurement of the RV and LV diameters, a perpendicular axis of
the ventricular cavities from the endocardium to the interventricular
septum on a standard axial image was measured, using the widest
diameter for each chamber. In cases where the measurements of
PA/AA and RV/LV ratios were discordant, a consensus read was
conducted with both radiologists present. A discordant result was
dened as a result where the readers’ measured values resulted in a
PA/AA or RV/LV ratio that was on opposing sides of the cut-points
dened above. In these cases, a third consensus measurement was
taken, and the final measurement was determined by averaging
all three readings. This collaborative approach aimed to resolve
discrepancies and improve measurement accuracy.
Echocardiography
Transthoracic echocardiography was used to evaluate cardiac
function. Standard 2D and Doppler echocardiography was performed
by a trained and certied echocardiography technologist. e ESC
guidelines were followed to assess variables including systolic le
ventricular function (Simpson biplane method), peak velocities of
Ewave and A wave, E/A ratio, right atrial pressure, tricuspid annular
AJTCCM VOL. 31 NO. 1 2025 21
ORIGINAL ARTICLES: RESEARCH
plane systolic excursion, inferior vena cava
diameter (inspiration and expiration), right
ventricular ejection fraction, TRVmax and
RVS P.
Spirometry
A trained technologist conducted spirometry
based on standardised American Thoracic
Society/ERS criteria.[16] Low forced vital
capacity (FVC) was dened as values <80%
predicted using the Global Lung Function
Initiative 2012 reference ranges or below the
lower limit of normal.[16]
Statistical analysis
Baseline data were presented as mean values
along with their standard deviations (SDs)
for normally distributed variables. Spearman
correlation coefficients were calculated to
assess the associations between PAD, RV/
LV ratio, PA/AA ratio, RVSP and TRVmax.
To measure the dierences in PA/AA ratio,
PAD and spirometric data (forced expiratory
volume in 1 second (FEV1), FVC) between
groups with and without echocardiographic
evidence of PH, an independent t-test
was conducted, as PH was considered
unlikely in patients with an RVSP <36 and
a TRVmax <2.8. Further, a multiple linear
regression analysis was used to investigate the
correlations between CT chest image metrics
(specically PAD, PA/AA ratio and RV/LV
ratio), RVSP and TRVmax.
Results
We assessed the records of 173 patients who
had previously been treated for pulmonary
TB. Of these, 121 patients were excluded,
of whom 8 had left heart disease, 57 had
alternative known causes for PH, 16 had
echocardiography performed >9 months
aer the CT chest scan, and 40 had various
other reasons for exclusion; 52 patients were
therefore included (Fig.1).
Table1 summarises the baseline clinical
characteristics. Participants had a mean (SD)
age of 42.15 (12.60) years, with 27 males
(52%) and 25 females (48%). Eight (15%)
were HIV positive, 1 (2%) had diabetes
mellitus, and 5 (10%) had hypertension. Only
14 (30%) were non-smokers, with 17 (36%)
reporting current smoking. More than half
(52%) reported two or more episodes of TB.
e mean (SD) predicted FEV1 was 42.9%
(18.1%) and the mean FVC 59.6% (17.7%).
e 52 patients were separated into two
groups, PA/AA ratio >1 (n=26) and PA/AA
ratio ≤1 (n=26). ere were no signicant
dierences between the PA/AA ≤1 and >1
groups with regard to sex, smoking status,
lung function, presence of obesity, HIV,
hypertension, diabetes mellitus or number
of TB episodes. However, in the high PA/
AA group, there were significantly more
participants aged 14-34years (p=0.01).
Table 2 summarises the spirometric
results and echocardiographic findings.
Echocardiography-measured RVSP was
higher in the PA/AA >1 group than in the
group with a ratio ≤1 (mean (SD) 37.5 (30.9)
mmHg v. 19.8 (12.2) mmHg, respectively;
p=0.05). In 17 patients RVSP could not be
measured because no tricuspid regurgitant
jet could be seen, and in the absence of
alternative echocardiographic criteria, these
patients were assumed not to have PH. Five
(29%) of these patients had a PA/AA ratio >1.
The CT scan-measured PA/AA ratio
correlated linearly with RVSP and TRVmax.
A scatter plot displayed a clear positive
linear correlation (Spearman’s correlation
coefficient 0.4420 (p=0.0083) and 0.3444
(p=0.0582), respectively), indicating a
moderate positive correlation (Fig. 2).
Similarly, RVSP was greater in patients with
a raised PAD (p=0.007) and exhibited a
significant positive correlation with PAD
(Spearman’s correlation coefficient 0.5343
(p=0.0011)) (Fig.2). Additionally, TRVmax
showed a robust correlation with PAD
(Spearman’s correlation coefficient 0.5345
(p=0.0023)).
Patients without echocardiography-
measured PH were statistically more likely to
have a normal RV/LV ratio on CT scan than
those with PH (p<0.001). However, on further
analysis there was no direct correlation of CT
scan-measured RV/LV ratio with RVSP or
TRVmax (Spearman’s correlation coecient
0.2507 (p=0.1729) and 0.3486 (p=0.0749),
respectively), indicating at best a weak
positive correlation for TRVmax (Fig.2).
e PA/AA ratio was statistically greater
in patients with echocardiography-measured
PH compared with those without PH (PA/AA
1.25 v. 0.95, respectively; p<0.001). Similarly,
the PAD was significantly greater in the
group with PH than in the group without
echocardiographic evidence of PH (34.2 mm
v. 26.7 mm, respectively; p<0.001) (Fig.3A).
Patients without echocardiographic PH
had a 13.7% higher mean FVC (% predicted)
compared with those with PH (mean (SD)
61.3% (16.3%) v. 47.6% (18.3%), respectively;
p=0.044). is nding suggests a relationship
between the presence of PH and reduced
FVC. Interestingly, the same was not found
for FEV1, with no difference in FEV1 (%
predicted) found between those with and
without echocardiographic PH (p=0.1543)
(Fig.3B). ere was no signicant dierence
in either the FEV1 (% predicted) or FVC
(% predicted) between patients with and
Patients excluded,
n=121
• Left heart disease, n=8
• Concomitant pulmonary
pathology, n=55
• Concomitant liver diease, n=2
• Chest scan and echocardiography
performed >9 months apart, n=16
• Other reasons (missing data, unreliable
echocardiographic measurement
or poor views, CT scan not done or
deemed uninterpretable), n=40
Eligible patients,
n=52
Potential eligible patients,
N=173
Fig.1. Consort diagram of patients recruited. (CT = computed tomography.)
22 AJTCCM VOL. 31 NO. 1 2025
ORIGINAL ARTICLES: RESEARCH
without a raised PA/AA ratio (p=0.1367 and p=0.3889, respectively)
(Fig.3C).
Diagnostic tests demonstrated that a PA/AA ratio >1 had ~100%
sensitivity and 65% specificity for predicting the presence of
echocardiographic PH, with a positive predictive value (PPV) of
39.1% and a negative predictive value (NPV) of 100%.
e RV/LV ratio demonstrated 50% sensitivity, 97.2% specicity and
a PPV of 80%, while the NPV was high at 89.7%. PAD demonstrated
88.8% sensitivity and 42.5% specicity. e PPV was low at 25.8%,
and the NPV was 94.4%.
Discussion
We explored the utility of CT chest imaging as a complementary tool
to assist in the assessment of PH in patients with PTLD. We found
signicant correlations between the radiographic measurements of
the PA/AA ratio and PAD and the echocardiographic measures of
PH, namely RVSP and TRVmax. Patients with a raised PA/AA ratio
had a 17.7 mmHg higher RVSP. In our population, the sensitivity of a
PA/AA ratio >1 for detecting PH defined according to
echocardiographic criteria approached 100%; however, the specicity
(65%) and PPV (39.1%) were low, implying high false-positive rates.
Additionally, we found signicant correlations between some lung
function parameters and echocardiographically measured PH, but
not the PA/AA ratio.
PTLD oen occurs in settings with a lack of access to specialised
investigations, in particular echocardiography and RHC. A number of
CT scan measurements have been used to predict PH with reported
good sensitivity and specicity.[14,17] e PA/AA ratio is one such
measurement, and has been tested in other respiratory diseases. In
COPD, where echocardiographic measurement can be problematic
owing to lung hyperination and diculty in obtaining clear acoustic
windows, the PA/AA ratio can predict the presence of PH with 73%
sensitivity and 84% specicity.[12] However, in pulmonary brosis the
predictive value of this CT measurement appears inadequate.[18,19]
In PTLD, no data inform use of the PA/AA ratio to determine
the presence of PH.[1,6] Yet the PA/AA ratio is frequently reported
in clinical practice with extrapolations of interpretation from
other group3 PH-related diseases, despite potential dierences in
Table1. Baseline characteristics of all participants, and patients with PA/AA ratios >1 and ≤1
Characteristics* Overall (N=52), n (%)
PA/AA ratio ≤1
(n=26), n (%)
PA/AA ratio >1
(n=26), n (%) p-value
Age (years) (mean (SD) 42.15 (12.60))
14-34 11 (21) 1 (4) 10 (38) 0.01
35-44 20 (38) 11 (42) 9 (35)
45-54 11 (21) 6 (23) 5 (19)
55-74 10 (19) 8 (31) 2 (8)
Sex 0.17
Male 27 (52) 16 (62) 11 (42)
Female 25 (48) 10 (38) 15 (58)
Comorbidities
Hypertension 5 (10) 2 (8) 3 (12) 0.64
Diabetes mellitus 1 (2) 1 (4) 0 (0) 0.31
HIV 8 (15) 2 (8) 6 (23) 0.12
Obesity 2 (4) 1 (4) 1 (4) 1.00
Dyslipidaemia 3 (6) 3 (12) 0 0.074
Smoking status 0.98
Non-smoker 14/47 (30) 7/24 (29) 7/23 (30)
Previous smoker 16/47 (34) 8/24 (33) 8/23 (35)
Current smoker 17/47 (36) 9/24 (38) 8/23 (35)
Smoking amount (pack-years) 0.64
≤10 12/31 (39) 7/15 (47) 5/16 (31)
11-20 11/31 (35) 5/15 (33) 6/16 (38)
>20 8/31 (26) 3/15 (20) 5/16 (31)
Number of episodes of TB 0.20
1 25 (48) 12 (46) 13 (50)
2 13 (25) 4 (15) 9 (35)
3 8 (15) 6 (23) 2 (8)
4 3 (6) 2 (8) 1 (4)
5 1 (2) 0 (0) 1 (4)
>5 2 (4) 2 (8) 0
PA = pulmonary artery; AA = ascending aorta; SD = standard deviation; TB = tuberculosis.
*Percentages are calculated based on available data for each characteristic, as not all information was available for all participants. Denominators are provided for clarity when the totals for the
characteristic dier from the column totals.
AJTCCM VOL. 31 NO. 1 2025 23
ORIGINAL ARTICLES: RESEARCH
underlying pathology. In our population, we found that the PA/AA
ratio was helpful when low (≤1.0), and was able to exclude the presence
of PH, with an NPV approaching 100%. However, conversely, a raised
PA/AA ratio was not able to rule in the diagnosis of PH with certainty,
having a PPV of 39.1%. is nding implies that 6 out of 10 patients
with a raised PA/AA >1 will not have echocardiographic PH. In our
population of PTLD, a normal PA/AA ratio can therefore potentially
be used to rule out the need for further investigation of PH. e
imaging ndings illustrated in Fig.4 highlight increased PA/AA and
RV/LV ratios in patients with and without PH.
Table2. Spirometry results and echocardiographic ndings for all participants, and patients with PA/AA ratios >1 and ≤1
Characteristics* Overall (N=52), n (%)†
PA/AA ratio ≤1
(n=26), n (%)†
PA/AA ratio >1
(n=26), n (%)†p-value
Dyspnoea score 0.19
mMRC grade 1 12/40 (30) 8/20 (40) 4/20 (20)
mMRC grade 2 13/40 (33) 4/20 (20) 9/20 (45)
mMRC grade 3 15/40 (38) 8/20 (40) 7/20 (35)
mMRC grade 4 0 0 0
Spirometry
FVC (L), mean (SD) 2.2 (0.8) 2.3 (0.8) 2.1 (0.7) 0.39
FVC (%), mean (SD) 59.6 (17.7) 60.6 (18.1) 58.5 (17.8) 0.70
FEV1 (L), mean (SD) 1.5 (1.6) 1.4 (0.6) 1.6 (2.2) 0.63
FEV1 (%), mean (SD) 42.9 (18.0) 46.3 (20.8) 39.7 (14.5) 0.23
FEV1/FVC ratio, mean (SD) 59.6 (21.1) 60.0 (21.5) 59.3 (21.3) 0.91
Echocardiographic ndings
LVEF (%), mean (SD) 57.9 (5.7) 58.2 (3.4) 57.6 (7.4) 0.75
TRVmax (m/s), mean (SD) 2.2 (1.2) 1.8 (0.8) 2.4 (1.4) 0.16
RAP (mmHg), mean (SD) 5.7 (2.6) 5.0 (0.0) 6.4 (3.5) 0.083
RVSP (mmHg), mean (SD) 30.4 (26.4) 19.8 (12.2) 37.5 (30.9) 0.05
TAPSE (mm), mean (SD) 17.9 (3.3) 19.1 (2.7) 16.8 (3.5) 0.014
Tricuspid valve regurgitation absent 30 (58) 18 (60) 12 (40) 0.092
Tricuspid valve regurgitation present 22 (42) 8 (36) 14 (64)
Pulmonary valve regurgitation absent 41/51 (80) 23/26 (88) 18/25 (72) 0.14
Pulmonary valve regurgitation present 10/51 (20) 3/26 (12) 7/25 (28)
IVC ndings
IVC size (mm), mean (SD) 15.8 (4.2) 15.8 (3.5) 15.8 (4.6) 0.98
IVC collapse >50% 36/45 (80) 18/24 (75) 18/21 (86) 0.032
IVC dilated 3/45 (7) 0 3/21 (14)
IVC not measured 4/35 (9) 4/24 (17) 0
IVC not visualised 2/45 (4) 2/24 (8) 0
RA size
Normal size 40/46 (87) 23/23 (100) 17/23 (74) 0.009
Dilated 6/46 (13) 0 6/23 (26)
RV function
RVEF normal 37/44 (84) 22/23 (96) 15/21 (71) 0.028
RVEF abnormal 7/44 (16) 1/23 (4) 6/21 (29)
RV size
Normal size 40/47 (85) 23/24 (96) 17/23 (74) 0.035
Dilated 7/47 (15) 1/24 (4) 6/23 (26)
Echocardiographic evidence of PH <0.001
No 40 (77) 26 (100) 14 (56)
Ye s 9 (17) 0 9 (36)
Unclear 3 (6) 0 3 (8)
PA = pulmonary artery; AA = ascending aorta; mMRC = Modied Medical Research Council; FVC = forced vital capacity; SD = standard deviation; FEV = forced expiratory volume in 1 second;
LVEF = le ventricular ejection fraction; TRVmax = maximum tricuspid regurgitant velocity; RAP = right atrial pressure; RVSP = right ventricular systolic pressure;
TAPSE = tricuspid annular plane systolic excursion; IVC = inferior vena cava; RA = right atrium; RV = right ventricle; RVEF = right ventricular ejection fraction; PH = pulmonary hypertension.
*Percentages are calculated based on available data for each characteristic, as not all information was available for all participants. Denominators are provided for clarity when the totals for the
characteristic dier from the column totals.
†Except where otherwise indicated.
24 AJTCCM VOL. 31 NO. 1 2025
ORIGINAL ARTICLES: RESEARCH
1.61.41.21
PA/AA ratio
0.80.6
TRVmax (m/s)
5
4
3
2
1
P=0.0083
1.61.41.21
PA/AA ratio
0.8
0.6
RVSP (mmHg)
100
50
0
P=0.00582
45403530
PAD (mm)
2520
RVSP (mmHg)
100
50
0
P=0.007
2.521.5
RV/LV ratio
10.5
TRVmax (m/s)
5
4
3
2
1
P=0.0749
2.521.5
RV/LV ratio
10.5
RVSP (mmHg)
100
50
0
P=0.1729
Fig.2. Correlation relationships between PAD, RV/LV ratio and PA/AA ratio on CT scan and TRV/max on echocardiography. (PAD = pulmonary artery diameter; RV = right ventricle; LV = le
ventricle; PA = pulmonary artery; AA = ascending aorta; CT = computed tomography; TRVmax = peak tricuspid regurgitant jet velocity.)
AJTCCM VOL. 31 NO. 1 2025 25
ORIGINAL ARTICLES: RESEARCH
No PH
PA/AA ratio
1.6
1.4
1.2
1
0.8
0.6
p<0.001A
PH No PH
FVC (% predicted)
100
80
60
40
20
p=0.044B
PH FVC <80% predicted
PA/AA ratio
1.4
1.2
1
0.8
0.6
p=0.3889 C
FVC >80% predicted
p<0.001
No PH
PAD (mm)
45
40
35
30
25
20
PH No PH
FEV1 (% predicted)
100
80
60
40
20
p=0.1543
PH PA/AA ratio <1
FEV1 (% predicted)
100
80
60
40
20
p=0.1367
PA/AA ratio >1
Fig.3. Box-and-whisker plots comparing (A) PA/AA ratio and PAD in patients with and without PA, (B) FVC and FEV1 predicted in patients with and without PH, and (C) FVC and FEV1 predicted
in patients with and without a raised PA/AA ratio. (PA = pulmonary artery; AA = ascending aorta; PAD = pulmonary artery diameter; FVC = forced vital capacity; FEV1 = forced expiratory
volume in 1 second; PH = pulmonary hypertension.)
26 AJTCCM VOL. 31 NO. 1 2025
ORIGINAL ARTICLES: RESEARCH
e reason for the high false-positive rate of the PA/AA ratio
is not clear. It is possible that the brotic destruction of the lung
parenchyma adjacent to the pulmonary artery may play a role and
cause distortion and dilation of the main pulmonary artery, in the
absence of elevated pulmonary artery pressures. Disappointingly,
we found no association between spirometric values and PA/AA
that would have supported this hypothesis. is lack of association
is likely to point to mixed causes of raised PA/AA, being due to
raised pressures in some cases and anatomical defects in others, and
requires further research. Interestingly, we did nd an association
between FVC and the presence of PH on echocardiography. is
is in contrast to our previous study, which showed no association
between spirometric findings and the presence of PH in a
non-healthcare-seeking post-TB population.
[20]
e mechanism for
PH aer TB is unclear and may include lung brosis with loss of
the capillary bed, a vasculopathy, increased thrombosis, and even
mediastinal brosis.
[6,8
]
Other CT scan measurements have been assessed in the context of
PH. Measurement of the RV/LV ratio in a general patient population
predicts PH with 85.7% sensitivity and 86.1% specicity,[21,22] but
usually requires cardiac gated images, which were not performed
in our study. In interstitial lung disease, an increased RV/LV ratio
reportedly predicts the presence of PH with 58% sensitivity and 70%
specicity and is associated with increased mortality.[23,24]
In our study, the RV/LV ratio had a low sensitivity of 50%, but a
specicity of 97.2%. e improved specicity compared with PA/AA
is noted and could be helpful to rule in the diagnosis of PH, with a
false-positive rate of 20% (PPV 80%).
Studies reported an average sensitivity and specicity for PH based
on a dilated PA as 71.9% and 81.1%, respectively.[22,25] In our cohort,
the PAD had sensitivity of 88.8%; however, the specicity of 42.5% and
PPV of 25.8% were lower, implying high false-positive rates.
From our data, it appears that of the three CT measurements
assessed, the PA/AA ratio may be the best screening measurement
for excluding the diagnosis of PH in PTLD, but at the cost of a high
false-positive rate.
These findings are potentially important in screening for PH in
PTLD, as PTLD occurs more frequently in low- and middle-income
countries, where access to echocardiography and RHC is limited, and
CT scan imaging is marginally more available. Furthermore, CT scan
measurements are less operator dependent than echocardiographic
measures, and do not require the presence of good acoustic windows,
which can be challenging in some patients with PTLD. Although
CT imaging cannot replace either echocardiography or RHC in the
diagnosis and management of PH, our data show that it may be helpful
in ruling out PH and reducing the number of referrals for further
investigation.
Study limitations and future directions
Limitations of this study include the small sample size, the retrospective
design and the possibility of selection bias. Further, we acknowledge
the absence of RHC to conrm PH diagnosis, which is considered the
A
C
B
D
Fig.4. A 50-year-old woman with one episode of previous TB has PH with a TRVmax of 3.52 m/s and an increased PA/AA ratio (A) and an
increased RV/LV ratio (B). A 40-year-old woman with one episode of previous TB but no PH has a TRVmax of 1.85 m/s but an increased
PA/AA ratio (C) and an increased RV/LV ratio (D). (TB = tuberculosis; PH = pulmonary hypertension; TRVmax = peak tricuspid regurgitant jet
velocity; PA = pulmonary artery; AA = ascending aorta; RV = right ventricle; LV = le ventricle.)
AJTCCM VOL. 31 NO. 1 2025 27
ORIGINAL ARTICLES: RESEARCH
gold standard test. Additionally, ECG gating on CT scanning, which
can minimise motion artifacts, was not routinely performed; however,
it is not performed on the majority of CT scans in settings where
PTLD is prevalent, and therefore allows a real-world comparison.
Conclusions
is study highlights the potential of CT chest imaging, particularly
the PA/AA ratio, as a tool for excluding signicant PH in patients
with PTLD. However, a raised PA/AA ratio will not be associated with
echocardiographic PH in 60% of cases, and other mechanisms causing
an increased PA/AA must not be forgotten in PTLD. ese ndings
suggest that CT imaging may have an important role to play in limiting
referrals for echocardiography in healthcare-limited settings. Further
research with larger cohorts and prospective designs is recommended
to validate these correlations and explore treatment options for PTLD
patients with PH.
Data availability. e datasets generated and analysed during the present
study may be available from the corresponding author (MA) on reasonable
request, pending institutional review board approval.
Declaration. e research for this study was done in partial fullment
of the requirements for MA’s MMed (Int Med) degree at Stellenbosch
University.
Acknowledgements. We thank Prof. Alfred Musekiwa for his support and
guidance in statistical analysis, and the patients and sta of our respiratory
clinic for the important contributions they have made to this work.
Author contributions. MA, EHL, BA: conceptualised and designed the
study. NB, MA, HT: data collection. CA, SG-R: radiographic assessments.
MA, EHL, BWA: results analysis and manuscript preparation. All authors:
results review and manuscript review.
Funding. None.
Conicts of interest.None.
1. ienemann F, Dzudie A, Mocumbi AO, et al. e causes, treatment, and outcome
of pulmonary hypertension in Africa: Insights from the Pan African Pulmonary
Hypertension Cohort (PAPUCO) Registry. Int J Cardiol 2016;221:205-211. https://doi.
org/10.1016/j.ijcard.2016.06.242
2. Galiè N, Humbert M, Vachiery JL, et al. 2015 ESC/ERS guidelines for the diagnosis
and treatment of pulmonary hypertension. Eur Heart J 2016;37(1):67-119. https://doi.
org/10.1093/eurheartj/ehv317
3. Gillenwater S, Martinez F, Drexler I, et al. Role of cardiovascular CT in pulmonary
hypertension. Curr Radiol Rep 2020;8(8). https://doi.org/10.1007/s40134-020-00354-2
4. Subramanyam P, Abouzeid C, Groner LK. Multimodality imaging of pulmonary
hypertension. Curr Treat Options Cardiovasc Med 2020;22(11). https://doi.org/10.1007/
s11936-020-00843-9
5. Fukunaga R, Glaziou P, Harris JB, Date A, Floyd K, Kasaeva T. Epidemiology of
tuberculosis and progress toward meeting global targets – worldwide, 2019. MMWR
Morb Mortal Wkly Rep 2021;70(12):427-430. https://doi.org/10.15585/mmwr.mm7012a4
6. Allwood BW, Maarman GJ, Kyriakakis CG, Doubell AF. Post-pulmonary tuberculosis
complications in South Africa and a potential link with pulmonary hypertension:
Premise for clinical and scientic investigations. S Afr Med J 2018;108(7):529-529.
https://doi.org/10.7196/SAMJ.2018.v108i7.13359.
7. Bhattacharyya P, Saha D, Bhattacherjee PD, Das SK, Bhattacharyya PP, Dey R.
Tuberculosis associated pulmonary hypertension: The revelation of a clinical
observation. Lung India 2016;33(2):135-139. https://doi.org/10.4103/0970-
2113.177433
8. Montani D, Günther S, Dorfmüller P, et al. Pulmonary arterial hypertension.
Orphanet J Rare Dis 2013;8:97. https://doi.org/10.1186/1750-1172-8-97
9. Fakharian A, Hamidi N, Hosseinloo BH, et al. Correlation between the pulmonary
artery pressure measured in echocardiography and pulmonary artery diameter
in the CT-scan of patients suffering from interstitial lung disease. Tanaffos
2011;10(3):37-41.
10. Grünig E, Peacock AJ. Imaging the heart in pulmonary hypertension: An update.
Eur Respir Rev 2015;24(138):653-664. https://doi.org/10.1183/16000617.0058-2015
11. Frost A, Badesch D, Gibbs JSR, et al. Diagnosis of pulmonary hypertension. Eur
Respir J 2019;53(1):1801904. https://doi.org/10.1183/13993003.01904-2018
12. Iyer AS, Wells JM, Vishin S, Bhatt SP, Wille KM, Dranseld MT. CT scan-measured
pulmonary artery to aorta ratio and echocardiography for detecting pulmonary
hypertension in severe COPD. Chest 2014;145(4):824-832. https://doi.org/10.1378/
chest.13-1422
13. Truong QA, Bhatia HS, Szymonia J, et al. A four-tier classication system of
pulmonary artery metrics on computed tomography for the diagnosis and prognosis
of pulmonary hypertension. J Cardiovasc Comput Tomogr 2018;12(1):60-66.
https://doi.org/10.1016/j.jcct.2017.12.001
14. Ng CS, Wells AU, Padley SP. A CT sign of chronic pulmonary arterial hypertension:
The ratio of main pulmonary artery to aortic diameter. J Thorac Imaging
1999;14(4):270-278. https://doi.org/10.1097/00005382-199910000-00007
15. Greiner S, Jud A, Aurich M, et al. Reliability of noninvasive assessment of systolic
pulmonary artery pressure by Doppler echocardiography compared to right
heart catheterisation: Analysis in a large patient population. J Am Heart Assoc
2014;3(4):e001103. https://doi.org/10.1161/JAHA.114.001103
16. Quanjer PH, Stanojevic S, Cole TJ, et al. Multi-ethnic reference values for spirometry
for the 3–95-yr age range: e global lung function 2012 equations. Eur Respir J
2012;40(6):1324-1343. https://doi.org/10.1183/09031936.00080312
17. Mahammedi A, Oshmyansky A, Hassoun PM, Thiemann DR, Siegelman SS.
Pulmonary artery measurements in pulmonary hypertension. J orac Imaging
2013;28(2):96-103. https://doi.org/10.1097/RTI.0b013e318271c2eb
18. Zisman DA, Karlamangla AS, Ross DJ, et al. High-resolution chest CT ndings do not
predict the presence of pulmonary hypertension in advanced idiopathic pulmonary
brosis. Chest 2007;132(3):773-779. https://doi.org/10.1378/chest.07-0116
19. Devaraj A, Wells AU, Meister MG, Corte TJ, Hansell DM. e eect of diuse
pulmonary fibrosis on the reliability of CT signs of pulmonary hypertension.
Radiology 2008;249(3):1042-1049. https://doi.org/10.1148/radiol.2492080269
20. Louw E, Baines N, Maarman G, et al. e prevalence of pulmonary hypertension
aer successful tuberculosis treatment in a community sample of adult patients.
Pulm Circ 2023;13(1):e12184. https://doi.org/10.1002/pul2.12184
21. Chan AL, Juarez MM, Shelton DK, et al. Novel computed tomographic chest
metrics to detect pulmonary hypertension. BMC Med Imaging 2011;11:7. https://
doi.org/10.1186/1471-2342-11-7
22. Altschul E, Remy-Jardin M, Machnicki S, et al. Imaging of pulmonary hypertension.
Chest 2019;156(2):211-227. https://doi.org/10.1016/j.chest.2019.04.003
23. Bax S, Jacob J, Ahmed R, et al. Right ventricular to le ventricular ratio atCT
pulmonary angiogram predicts mortality in interstitial lung disease. Chest
2020;157(1):89-98. https://doi.org/10.1016/j.chest.2019.06.033
24. Bax S, Bredy C, Dimopoulos K, et al. Right ventricular dimensions on CTPA predict
pulmonary hypertension and prognosis in interstitial lung disease (ILD). ERS
International Congress 2017 abstracts, published online September 2017. https://
doi.org/10.1183/1393003.congress-2017.PA2433
25. Ascha M, Renapurkar R, Tonelli AR. A review of imaging modalities in pulmonary
hypertension. Ann orac Med 2017;12(2):61-73. https://doi.org/10.4103/1817-
1737.203742
Received 14 February 2024. Accepted 6 January 2025. Published 28 March 2025.
