1. CIFASD: Dysmorphology Core
Kenneth L. Jones
Ludmila Bakhireva
Luther K. Robinson
H. Eugene Hoyme
Miguel del Campo
Christina D. Chambers
June, 2005
Santa Barbara, CA
First of all, I would like to thank my collaborators: ……
who encouraged me to pursue this topic and provided valuable feedback during the entire process.
I also would like to thank my husband, Alexei Bakhirev, for his support, understanding, and willingness to listen to me as I worked my way through this project

2. Progress Report of the Dysmorphology Core (DC):
DC Exam Form: finalized
DC Manual for Pediatricians: finalized
Access Data Input Tool: field tested, modified accordingly:
Race/ethnicity fields included
Country of origin
Validation rules implemented
Data obtained from 6 sites (N=467 as of April 05)
Preliminary data analysis conducted
Osteoporosis and atherosclerosis are usually thought to be two independent processes that occur with aging. However, there is an increasing interest that these conditions are associated.
Links between the two conditions were first noted more than one hundred years ago by German Pathologist Rudolf Virchow who first reported morphologic similarities between calcified plaque and bone tissue.
Roentgenographic studies in the 1970s showed an association between calcification of the abdominal aorta and osteoporosis of the lumbar spine
Recent molecular biology studies discovered cells with both osteoblastic and osteoclastic potential in vascular tissue, and bone-related proteins have been identified in calcified arterial and valvular lesions.
Historically, this ectopic calcification was considered a degenerative process leading to passive precipitation of calcium phosphate. Molecular biology revealed that artery calcification is in fact an active process which is regulated by biological mechanisms similar to those of bone formation

3. Clinical Sites which Submitted Data to the Central Repository (N=467)
% from total
Luther Robinson, Buffalo 93
19.9
Phil May, Rome
213
45.6
Sarah Mattson, San Diego 23
4.9
Sarah Mattson, Moscow 50
10.7
Sarah Mattson, Finland 67
14.3
Sandra Jacobson, South Africa 21
4.3

4. Number of Children Examined by Each Pediatrician-Dysmorphologist
Examiner N
% from total
Hoyme
138
29.6
Jones 83
17.8
Khaole 21
4.5
Robinson
192
41.1
Del Campo 33
7.1

5. Description of the Sample (N=467)

6. Specific Aims:
Which cut-off point (≤10th % vs ≤3rd %) for weight, height, and OFC is better predictor of FAS.
Define the best cut-off point for PFL.
Examine whether PFL is independent from OFC measure.
Identify the prevalence of all key and additional structural features among children with FAS.

7. Specific Aim # 1: Proportion of Growth Deficient Children (among FAS-diagnosed) at Each Cut-off:

8. Specific Aim # 1 (cont’d): Sensitivity & Specificity of FAS Diagnosis at Each Cut-off for Each Parameter of Growth

9. Specific Aim # 1 (cont’d): Sensitivity & Specificity of FAS Diagnosis at Each Cut-off for Each Parameter of Growth

10. Specific Aim # 2 Define Best Cut-off Point for PFL
Research question:
Among children with FAS (N=113),
define a proportion with short PFL
at each cut-off point:
PFL≤3rd percentile: %
PFL ≤10th percentile: %
PFL ≤25th percentile: %

11. Specific Aim # 2 (cont’d) Define Best Cut-off Point for PFL
Research question:
Estimate the effectiveness of each cut-off point for PFL percentile (0-100)
as a continuous measure
to discriminate between
FAS-positive & FAS-negative children.

12. Specific Aim # 2 (cont’d) Define Best Cut-off Point for PFL
Methods:
Receiver Operating Characteristic (ROC) curve was used
ROC curve is the plot of sensitivity vs specificity of FAS diagnosis for all possible threshold values of PFL percentile

14. Specific Aim # 2 (cont’d) Results
PFL percentile of ’10’ has the best ability to differentiate between FAS+ & FAS- children
PFL%=10: Sensitivity=82%, Specificity=96%
PFL%=25 has too low specificity (75%)
PFL%=3 has too low sensitivity (48%)

15. Specific Aim # 3 Examine whether PFL is Independent from OFC Measure.
Research question:
Among FAS+ kids with normal OFC,
what is a proportion of children
with short PFL :
N= 24 (children w/ FAS who have normal OFC)
PFL≤3rd percentile: %
PFL ≤10th percentile: %

16. Specific Aim # 3 (cont’d) Correlation b/w OFC & PFL (all children)
Correlation Coefficient (R) = (p=<0.0001)
R2 = 0.29
29% of the variation in PFL can be explained by variation in OFC
71% of the variance in PFL is due to factors other than OFC

17. Specific Aim # 3 (cont’d) Correlation b/w OFC & PFL (all children)

18. Specific Aim # 3 (cont’d) Correlation b/w OFC & PFL (FAS children)
Correlation Coefficient (R) = (p=<0.001)
R2 = 0.10
10% of the variation in PFL can be explained by variation in OFC
90% of the variance in PFL is due to factors other than OFC

19. Specific Aim # 3 (cont’d) Correlation b/w OFC & PFL (FAS children)

20. Key Structural Feature
Specific Aim # 4: Prevalence of Key Structural Features among Children Diagnosed with FAS
Key Structural Feature
Prevalence (%)
PFL≤10th percentile
81.1
Smooth philtrum
83.2
Thin Vermilion border
86.7

21. Additional Structural Feature
Specific Aim # 4 Sensitivities & Specificities of Additional Structural Features
Additional Structural Feature
Sensitivity (%)
Specificity (%)
Railroad track configuration of ears
10.6
97.7
Ptosis
13.3
98.1
Camptodactyly
40.7
94.4
Difficulty in pronation/supination of elbows
8.9
Contractures in other joints
4.4
98.6
Hockey stick crease
24.8
91.6
Other altered palmar creases
21.2
93.5
Heart murmur
99.5

22. Conclusions
For Weight, Height, OFC cut-off point at 10th percentile yields much higher sensitivity of FAS diagnosis with very insignificant loss in specificity compared to the 3rd percentile
Best cut-off point for PFL in diagnosing FAS is 10th percentile
Only 29% variability in PFL can be explained by OFC among all children & only 10% among children with FAS.

23. Conclusions (cont’d)
Among key structural features, thin vermilion border has the highest prevalence among children with FAS.
Among additional structural features, camptodactyly occurs with the highest frequency among children with FAS diagnosis followed by altered palmar creases