Abstract: TH-PO832
Performance of Ultrasound-Derived Average Kidney Length (aKL) for Risk Prediction in Autosomal Dominant Polycystic Kidney Disease (ADPKD)
Session Information
- Cystic Kidney Diseases: Clinical
November 07, 2019 | Location: Exhibit Hall, Walter E. Washington Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: Genetic Diseases of the Kidneys
- 1001 Genetic Diseases of the Kidneys: Cystic
Authors
- Akbari, Pedram, Toronto General Hospital UHN, Toronto, Ontario, Canada
- Nasri, Fatemeh, Toronto General Hospital UHN, Toronto, Ontario, Canada
- Quist, Crystal F., Toronto General Hospital UHN, Toronto, Ontario, Canada
- Guiard, Elsa, Toronto General Hospital UHN, Toronto, Ontario, Canada
- Iliuta, Ioan-Andrei, University Health Network and University of Toronto, Toronto, Ontario, Canada
- Ahmed, Syed Essam, Toronto General Hospital UHN, Toronto, Ontario, Canada
- Calvaruso, Luca, University Health Network, Toronto, Canada
- Atri, Mostafa, University Health Network and University of Toronto, Toronto, Ontario, Canada
- Khalili, Korosh, University Health Network and University of Toronto, Toronto, Ontario, Canada
- Pei, York P., University Health Network and University of Toronto, Toronto, Ontario, Canada
Background
Total kidney volume (TKV) is a validated prognostic biomarker for risk assessment in ADPKD. However, accurate TKV measurements require magnetic resonance imaging (MRI), which is relatively expensive, time-consuming, and not readily accessible. To circumvent these issues, an ultrasound-derived aKL >16.5 cm (Kidney Int 88: 146-151, 2015) has been proposed as a diagnostic thresohld to identify patients with ADPKD at high-risk for progression. However, this approach has not been tested in a real world patient cohort.
Methods
We conducted a prospective study of 123 patients recruited from a PKD specialty center who underwent a standardized 3D-US and MRI. KLs from 3D-US were measured by 5 different experienced technicians; KLs and TKVs from MRI were derived by an experienced radiologist blinded to patient clinical results. Bland-Altman plots were used to assess the agreement of KLs by US vs. MRI. TKV adjusted for age and height was used to derive the Mayo Clinic Imaging Class (MCIC) and used as a "gold standard" for classifiying patients into low-risk (1A-1B) vs. high-risk (1C-1E) grouping.
Results
Table 1 shows the study patient characteristics. Good agreement was observed between US- and MRI-derived aKL with 98% of cases within 20% difference. Using aKL >16.5 cm as a diagnostic threshold yielded a sensitivity of 0.53, specificity of 0.92, false positive rate (FPR) of 0.14, false negative rate (FNR) of 0.31, and accuaracy of 0.74 for predicting high-risk MCIC (1C-1E).
Conclusion
US-derived aKL >16.5 cm provides a simple approach for risk stratification in ADPKD, but is associated with a FPR of 0.14 (i.e. 14% of low-risk patients misclassified as high-risk) and FNR of 0.31 (i.e. 31% of high-risk patients misclassified as low-risk). When a therapeutic decision (i.e. use of Tolvaptan) is based on accurate risk stratefication these errors have clinical consequence.
Table 1. Characteristics of Patient Cohort
Total (n=123) | |
Age (years) mean ± SD | 45 ± 15 |
Gender, M:F | 1 : 1.2 |
s-Cr (mg/dL) mean ± SD | 1.14 ± 0.48 |
Height (meters) mean ± SD | 1.69 ± 0.9 |
TKV (mL) mean ± SD | 1116 ± 822 |
Funding
- Government Support - Non-U.S.