Predictive Models of Emphysema Progression in...
Transcript of Predictive Models of Emphysema Progression in...
![Page 1: Predictive Models of Emphysema Progression in HIVregist2.virology-education.com/2015/6hivaging/16_Leung.pdf0.0003 0.560 ASCVD (Cardiovascular Risk Score) 0.002 0.004 0.661 Multivariable](https://reader031.fdocuments.in/reader031/viewer/2022011800/5ad554b37f8b9a5c638cf681/html5/thumbnails/1.jpg)
Predictive Models of Emphysema Progression in HIV
Janice Leung, MD FRCPCA. Malagoli, A. Santoro, G. Besutti, G.
Ligabue, D. Dai, J. Mayo, S. Lam, D. Sin, S. Man, G. Guaraldi
![Page 2: Predictive Models of Emphysema Progression in HIVregist2.virology-education.com/2015/6hivaging/16_Leung.pdf0.0003 0.560 ASCVD (Cardiovascular Risk Score) 0.002 0.004 0.661 Multivariable](https://reader031.fdocuments.in/reader031/viewer/2022011800/5ad554b37f8b9a5c638cf681/html5/thumbnails/2.jpg)
COPD: An Emerging Comorbidity
• Incidence rates of COPD appear to be higher in HIV compared to non-HIV patients
Crothers K. Am J Respir Crit Care Med 2011.
0
5
10
15
20
25
30
35
40
<40 40-44 45-49 50-60 ≥60
HIV+HIV-In
cide
nce
(per
1,0
00 p
erso
n-ye
ars)
Age (Years)
p-trend<0.001
![Page 3: Predictive Models of Emphysema Progression in HIVregist2.virology-education.com/2015/6hivaging/16_Leung.pdf0.0003 0.560 ASCVD (Cardiovascular Risk Score) 0.002 0.004 0.661 Multivariable](https://reader031.fdocuments.in/reader031/viewer/2022011800/5ad554b37f8b9a5c638cf681/html5/thumbnails/3.jpg)
Faster Lung Function Decline
Drummond MB. AIDS 2013.
FEV
1 (L
)
Duration of Follow-Up (Years)
2,6
2,8
3
3,2
3,4
0 1 2 3
HIV-HIV+/VL<70,000 copies/mLHIV+/VL>70,000 copies/mL
p=0.004
![Page 4: Predictive Models of Emphysema Progression in HIVregist2.virology-education.com/2015/6hivaging/16_Leung.pdf0.0003 0.560 ASCVD (Cardiovascular Risk Score) 0.002 0.004 0.661 Multivariable](https://reader031.fdocuments.in/reader031/viewer/2022011800/5ad554b37f8b9a5c638cf681/html5/thumbnails/4.jpg)
Worse Respiratory Symptoms
.Leung JM. AIDS Patient Care STDS 2014.
0
10
20
30
40
50
60
70
80
90
100
20 40 60 80 100 120 140
FEV1 % Predicted
SGR
Q T
otal
Sco
re
![Page 5: Predictive Models of Emphysema Progression in HIVregist2.virology-education.com/2015/6hivaging/16_Leung.pdf0.0003 0.560 ASCVD (Cardiovascular Risk Score) 0.002 0.004 0.661 Multivariable](https://reader031.fdocuments.in/reader031/viewer/2022011800/5ad554b37f8b9a5c638cf681/html5/thumbnails/5.jpg)
The Emphysema Phenotype in HIV
0
20
40
60
80
100
120
FEV1 (% Pred)FVC (% Pred)FEV1/FVC (%)
Emphysema Severity
Perc
ent
Guaraldi G. PLoS ONE 2014.
![Page 6: Predictive Models of Emphysema Progression in HIVregist2.virology-education.com/2015/6hivaging/16_Leung.pdf0.0003 0.560 ASCVD (Cardiovascular Risk Score) 0.002 0.004 0.661 Multivariable](https://reader031.fdocuments.in/reader031/viewer/2022011800/5ad554b37f8b9a5c638cf681/html5/thumbnails/6.jpg)
Clinical Significance of Emphysema
• >25% emphysema is significantly associated with a higher MRC dyspnea score, worse 6 minute walk distance, and higher BODE index in COPD1
• Emphysema is also significantly associated with worse St. George’s Respiratory Questionnaire scores2
1Diaz AA. Chest 2010.2Gietema HA. Respir Med 2013.
![Page 7: Predictive Models of Emphysema Progression in HIVregist2.virology-education.com/2015/6hivaging/16_Leung.pdf0.0003 0.560 ASCVD (Cardiovascular Risk Score) 0.002 0.004 0.661 Multivariable](https://reader031.fdocuments.in/reader031/viewer/2022011800/5ad554b37f8b9a5c638cf681/html5/thumbnails/7.jpg)
Study Objective
To determine the predictors of emphysema progression over a five year period in HIV and compare the
prevalence of progression to a non-HIV population
![Page 8: Predictive Models of Emphysema Progression in HIVregist2.virology-education.com/2015/6hivaging/16_Leung.pdf0.0003 0.560 ASCVD (Cardiovascular Risk Score) 0.002 0.004 0.661 Multivariable](https://reader031.fdocuments.in/reader031/viewer/2022011800/5ad554b37f8b9a5c638cf681/html5/thumbnails/8.jpg)
Methods
• HIV Study Population 345 HIV+ patients with ≥2 chest CT scans
Enrolled in an HIV Metabolic Clinic in Modena, Italy
Median Time Span (IQR) Between First and Last CT scans
= 3.36 years (2.07, 4.99)
![Page 9: Predictive Models of Emphysema Progression in HIVregist2.virology-education.com/2015/6hivaging/16_Leung.pdf0.0003 0.560 ASCVD (Cardiovascular Risk Score) 0.002 0.004 0.661 Multivariable](https://reader031.fdocuments.in/reader031/viewer/2022011800/5ad554b37f8b9a5c638cf681/html5/thumbnails/9.jpg)
Qualitative Emphysema Scoring
0 1=1-25% 2=26-50%
3=51-75% 4=76-100%
![Page 10: Predictive Models of Emphysema Progression in HIVregist2.virology-education.com/2015/6hivaging/16_Leung.pdf0.0003 0.560 ASCVD (Cardiovascular Risk Score) 0.002 0.004 0.661 Multivariable](https://reader031.fdocuments.in/reader031/viewer/2022011800/5ad554b37f8b9a5c638cf681/html5/thumbnails/10.jpg)
Qualitative Emphysema Scoring
Total Emphysema Score = NRUL + NRML + NRLL + NLUL + NLingula + NLLL
Emphysema ProgressionDefined as an increase in total emphysema score over the study period
![Page 11: Predictive Models of Emphysema Progression in HIVregist2.virology-education.com/2015/6hivaging/16_Leung.pdf0.0003 0.560 ASCVD (Cardiovascular Risk Score) 0.002 0.004 0.661 Multivariable](https://reader031.fdocuments.in/reader031/viewer/2022011800/5ad554b37f8b9a5c638cf681/html5/thumbnails/11.jpg)
Emphysema Distribution
Paraseptal Centrilobular
![Page 12: Predictive Models of Emphysema Progression in HIVregist2.virology-education.com/2015/6hivaging/16_Leung.pdf0.0003 0.560 ASCVD (Cardiovascular Risk Score) 0.002 0.004 0.661 Multivariable](https://reader031.fdocuments.in/reader031/viewer/2022011800/5ad554b37f8b9a5c638cf681/html5/thumbnails/12.jpg)
Emphysema Distribution
Centrilobular and Paraseptal
![Page 13: Predictive Models of Emphysema Progression in HIVregist2.virology-education.com/2015/6hivaging/16_Leung.pdf0.0003 0.560 ASCVD (Cardiovascular Risk Score) 0.002 0.004 0.661 Multivariable](https://reader031.fdocuments.in/reader031/viewer/2022011800/5ad554b37f8b9a5c638cf681/html5/thumbnails/13.jpg)
Baseline Characteristics (n=345)
Variable Median (IQR) or n (%)Age 49 (45, 53)Male Sex (%) 311 (90%)BMI 23.8 (21.9, 26.0)Current CD4 577 (430, 746)Nadir CD4 188 (65, 290)Viral Load <40 267 (77%)Smoking Status
NeverFormerCurrent, <10 cig/dayCurrent, ≥10 cig/dayN/A
64 (19%)101 (29%)70 (20%)96 (28%)14 (4%)
Smoking Pack-Years 17 (5, 28)
![Page 14: Predictive Models of Emphysema Progression in HIVregist2.virology-education.com/2015/6hivaging/16_Leung.pdf0.0003 0.560 ASCVD (Cardiovascular Risk Score) 0.002 0.004 0.661 Multivariable](https://reader031.fdocuments.in/reader031/viewer/2022011800/5ad554b37f8b9a5c638cf681/html5/thumbnails/14.jpg)
Baseline Respiratory Characteristics
Variable Median (IQR)FEV1/FVC 79.3 (75.6, 83.0)FEV1 (% Pred) 106.1 (96.7, 116.8)FVC (% Pred) 109.0 (98.3, 118.0)TLC (% Pred) 114.6 (106.1, 122.9)DLCO (% Pred) 75.2 (65.7, 87.2)Baseline Emphysema Score 0 (0, 3)
![Page 15: Predictive Models of Emphysema Progression in HIVregist2.virology-education.com/2015/6hivaging/16_Leung.pdf0.0003 0.560 ASCVD (Cardiovascular Risk Score) 0.002 0.004 0.661 Multivariable](https://reader031.fdocuments.in/reader031/viewer/2022011800/5ad554b37f8b9a5c638cf681/html5/thumbnails/15.jpg)
Variable Progressor(n=60)
Non-Progressor(n=285) P-value
Emphysema Score 4 (1, 6) 0 (0, 2) 7.81e-14
Emphysema TypeCentrilobularParaseptalCombination
3 (5%)9 (15%)37 (62%)
40 (14%)34 (12%)41 (14%)
1.13e-13
Smoking StatusNever or Former<10 cigs/day≥10 cigs/dayNA
15 (25%)11 (18%)32 (53%)2 (3%)
150 (53%)59 (21%)64 (22%)12 (4%)
8.89e-06
Smoking Pack-Years 22.5 (15.3, 31.8) 15.0 (1.0, 26.0) 0.0001
DLCO (%Pred) 64.75 (54.27, 78.08) 77.50 (68.15, 88.00) 0.0004
FEV1/FVC (%) 77.05 (72.44, 80.39) 79.78 (76.42, 83.55) 0.0007
BMI 22.93 (20.73, 24.61) 23.96 (21.98, 26.21) 0.0107
Male Sex 59 (98%) 252 (88%) 0.0159
Albumin 4.6 (4.3, 4.8) 4.7 (4.5, 4.9) 0.0379
![Page 16: Predictive Models of Emphysema Progression in HIVregist2.virology-education.com/2015/6hivaging/16_Leung.pdf0.0003 0.560 ASCVD (Cardiovascular Risk Score) 0.002 0.004 0.661 Multivariable](https://reader031.fdocuments.in/reader031/viewer/2022011800/5ad554b37f8b9a5c638cf681/html5/thumbnails/16.jpg)
Parameter Estimate Std. Error P-value
(Intercept) 0.740 0.308 0.017
Baseline DLCO (% Pred) -0.004 0.002 0.0005
Centrilobular and Paraseptal Distribution 0.241 0.097 0.013
Male Sex 0.100 0.078 0.200
Total Exposure to Integrase Inhibitors(Months)
-0.002 0.002 0.255
BMI -0.007 0.007 0.277
Baseline Emphysema Score 0.017 0.018 0.332
Albumin -0.335 0.349 0.338
Current Smoker ≥10 cigarettes/day -0.054 0.072 0.453
Duration of HIV (Months) -0.0002 0.0003 0.560
ASCVD (Cardiovascular Risk Score) 0.002 0.004 0.661
Multivariable Logistic Regression Model
Outcome Variable: Emphysema Progression
![Page 17: Predictive Models of Emphysema Progression in HIVregist2.virology-education.com/2015/6hivaging/16_Leung.pdf0.0003 0.560 ASCVD (Cardiovascular Risk Score) 0.002 0.004 0.661 Multivariable](https://reader031.fdocuments.in/reader031/viewer/2022011800/5ad554b37f8b9a5c638cf681/html5/thumbnails/17.jpg)
Line Model AUC
Full Model 0.911
Emphysema Score + Distribution + DLCO 0.847
Baseline FEV1/FVC 0.645
Baseline FEV1 %Pred 0.535
![Page 18: Predictive Models of Emphysema Progression in HIVregist2.virology-education.com/2015/6hivaging/16_Leung.pdf0.0003 0.560 ASCVD (Cardiovascular Risk Score) 0.002 0.004 0.661 Multivariable](https://reader031.fdocuments.in/reader031/viewer/2022011800/5ad554b37f8b9a5c638cf681/html5/thumbnails/18.jpg)
• Non-HIV Study Population 1,295 HIV- patients with ≥2 chest CT scans Enrolled in the Pan-Canadian Early
Detection of Lung Cancer Study
Comparisons to Non-HIV Population
• Qualitative Emphysema Scoring Total emphysema score
1 = <5% (Trivial)2 = 5-25% (Mild)3 = 26-50% (Moderate)4 = 51-75% (Severe)5 = 76-100% (Very Severe)
![Page 19: Predictive Models of Emphysema Progression in HIVregist2.virology-education.com/2015/6hivaging/16_Leung.pdf0.0003 0.560 ASCVD (Cardiovascular Risk Score) 0.002 0.004 0.661 Multivariable](https://reader031.fdocuments.in/reader031/viewer/2022011800/5ad554b37f8b9a5c638cf681/html5/thumbnails/19.jpg)
Matched Cohorts
Variable HIV (n=301) Non-HIV (n=301) P-value
Age (years) 49 (45, 54) 63 (60, 67) 2.2e-16
Male Sex 271 (90%) 271 (90%) 1
BMI (kg/m2) 23.9 (21.9, 26.1) 25.4 (23.5, 27.9) 1.269e-09
Smoking StatusCurrentFormerNever
142 (47%)97 (32%)62 (21%)
202 (67%)99 (33%)0 (0%)
2.2e-16
Smoking Pack-Years 17.0 (4.3, 26.0) 50.0 (41.3, 61.3) 2.2e-16
Maximum Time Interval (years) 3.01 (2.03, 4.24) 3.01 (2.03, 4.23) 0.9865
Emphysema Progressor 46 (15%) 52 (17%) 0.5809
![Page 20: Predictive Models of Emphysema Progression in HIVregist2.virology-education.com/2015/6hivaging/16_Leung.pdf0.0003 0.560 ASCVD (Cardiovascular Risk Score) 0.002 0.004 0.661 Multivariable](https://reader031.fdocuments.in/reader031/viewer/2022011800/5ad554b37f8b9a5c638cf681/html5/thumbnails/20.jpg)
Limitations
• Slight differences in emphysema scoring technique may limit comparisons to the non-HIV population.
• What is the appropriate HIV-negative control population?
• The practicalities of prediction model validation
![Page 21: Predictive Models of Emphysema Progression in HIVregist2.virology-education.com/2015/6hivaging/16_Leung.pdf0.0003 0.560 ASCVD (Cardiovascular Risk Score) 0.002 0.004 0.661 Multivariable](https://reader031.fdocuments.in/reader031/viewer/2022011800/5ad554b37f8b9a5c638cf681/html5/thumbnails/21.jpg)
Conclusions
• Emphysema progression in HIV can be predicted by baseline severity, distribution, and diffusion capacity
• These may be useful tools for the HIV clinician to identify high-risk patients in need of aggressive smoking cessation programs and COPD treatments
• Even younger HIV-infected patients with lesser smoking habits exhibit a similar prevalence of emphysema progression compared to HIV-uninfected patients
![Page 22: Predictive Models of Emphysema Progression in HIVregist2.virology-education.com/2015/6hivaging/16_Leung.pdf0.0003 0.560 ASCVD (Cardiovascular Risk Score) 0.002 0.004 0.661 Multivariable](https://reader031.fdocuments.in/reader031/viewer/2022011800/5ad554b37f8b9a5c638cf681/html5/thumbnails/22.jpg)
Future Directions
• Validation of predictive model in additional HIV cohorts with available CT imaging
• Determining the impact of emphysema progression on morbidity and mortality in HIV
• Identifying additional blood or lung-based biomarkers for emphysema progression in HIV
• Using next generation sequencing in lung tissue to identify the molecular determinants of emphysema progression in HIV
![Page 23: Predictive Models of Emphysema Progression in HIVregist2.virology-education.com/2015/6hivaging/16_Leung.pdf0.0003 0.560 ASCVD (Cardiovascular Risk Score) 0.002 0.004 0.661 Multivariable](https://reader031.fdocuments.in/reader031/viewer/2022011800/5ad554b37f8b9a5c638cf681/html5/thumbnails/23.jpg)
Thank YouSin/Man LabDr. S. F. Paul ManDr. Don SinStella XuJoseph LiuSheena TamYeni OhDavid Ngan
BC Cancer AgencyDr. Stephen LamSukhinder Kattra
BCCfE in HIV/AIDSDr. Julio MontanerDr. Marianne HarrisDr. Silvia GuillemiDr. Aida Moghadam
UBC Division of Radiology and Diagnostic ImagingDr. John MayoDr. Jonathan LeipsicDr. Cameron HagueDr. Darra Murphy
Modena HIV Metabolic Clinic, University of Modena and Reggio EmiliaDr. Giovanni GuaraldiDr. Andrea MalagoliDr. Antonella SantoroDr. Giulia BesuttiDr. Guido Ligabue
Proof Centre for ExcellenceDarlene Dai