Extending statistical learning for aneurysm rupture assessment to Finnish and Japanese populations using morphology, hemodynamics, and patient characteristics

Felicitas J. Detmer, Sara Hadad, Bong Jae Chung, Fernando Mut, Martin Slawski, Norman Juchler, Vartan Kurtcuoglu, Sven Hirsch, Philippe Bijlenga, Yuya Uchiyama, Soichiro Fujimura, Makoto Yamamoto, Yuichi Murayama, Hiroyuki Takao, Timo Koivisto, Juhana Frösen, Juan R. Cebral

Research output: Contribution to journalArticleResearchpeer-review

Abstract

OBJECTIVE: Incidental aneurysms pose a challenge for physicians, who need to weigh the rupture risk against the risks associated with treatment and its complications. A statistical model could potentially support such treatment decisions. A recently developed aneurysm rupture probability model performed well in the US data used for model training and in data from two European cohorts for external validation. Because Japanese and Finnish patients are known to have a higher aneurysm rupture risk, the authors' goals in the present study were to evaluate this model using data from Japanese and Finnish patients and to compare it with new models trained with Finnish and Japanese data. METHODS: Patient and image data on 2129 aneurysms in 1472 patients were used. Of these aneurysm cases, 1631 had been collected mainly from US hospitals, 249 from European (other than Finnish) hospitals, 147 from Japanese hospitals, and 102 from Finnish hospitals. Computational fluid dynamics simulations and shape analyses were conducted to quantitatively characterize each aneurysm's shape and hemodynamics. Next, the previously developed model's discrimination was evaluated using the Finnish and Japanese data in terms of the area under the receiver operating characteristic curve (AUC). Models with and without interaction terms between patient population and aneurysm characteristics were trained and evaluated including data from all four cohorts obtained by repeatedly randomly splitting the data into training and test data. RESULTS: The US model's AUC was reduced to 0.70 and 0.72, respectively, in the Finnish and Japanese data compared to 0.82 and 0.86 in the European and US data. When training the model with Japanese and Finnish data, the average AUC increased only slightly for the Finnish sample (to 0.76 ± 0.16) and Finnish and Japanese cases combined (from 0.74 to 0.75 ± 0.14) and decreased for the Japanese data (to 0.66 ± 0.33). In models including interaction terms, the AUC in the Finnish and Japanese data combined increased significantly to 0.83 ± 0.10. CONCLUSIONS: Developing an aneurysm rupture prediction model that applies to Japanese and Finnish aneurysms requires including data from these two cohorts for model training, as well as interaction terms between patient population and the other variables in the model. When including this information, the performance of such a model with Japanese and Finnish data is close to its performance with US or European data. These results suggest that population-specific differences determine how hemodynamics and shape associate with rupture risk in intracranial aneurysms.

Original languageEnglish
Pages (from-to)E16
JournalNeurosurgical focus
Volume47
Issue number1
DOIs
StatePublished - 1 Jul 2019

Fingerprint

Aneurysm
Rupture
Hemodynamics
Learning
Population
Area Under Curve
Intracranial Aneurysm
Statistical Models
Population Characteristics
Hydrodynamics
ROC Curve
Physicians
Therapeutics

Keywords

  • AUC = area under the receiver operating characteristic curve
  • BL = bulge location
  • cerebral aneurysm
  • CFD = computational fluid dynamics
  • hemodynamics
  • HWR = height/width ratio
  • IA = intracranial aneurysm
  • KE = kinetic energy
  • LSA = low shear area
  • MLN = mean surface curvature
  • morphology
  • NSI = nonsphericity index
  • OSImax = maximum oscillatory shear stress
  • risk
  • rupture
  • SAH = subarachnoid hemorrhage
  • WSS = wall shear stress

Cite this

Detmer, Felicitas J. ; Hadad, Sara ; Chung, Bong Jae ; Mut, Fernando ; Slawski, Martin ; Juchler, Norman ; Kurtcuoglu, Vartan ; Hirsch, Sven ; Bijlenga, Philippe ; Uchiyama, Yuya ; Fujimura, Soichiro ; Yamamoto, Makoto ; Murayama, Yuichi ; Takao, Hiroyuki ; Koivisto, Timo ; Frösen, Juhana ; Cebral, Juan R. / Extending statistical learning for aneurysm rupture assessment to Finnish and Japanese populations using morphology, hemodynamics, and patient characteristics. In: Neurosurgical focus. 2019 ; Vol. 47, No. 1. pp. E16.
@article{a3b831de2fff483485500f33c9912523,
title = "Extending statistical learning for aneurysm rupture assessment to Finnish and Japanese populations using morphology, hemodynamics, and patient characteristics",
abstract = "OBJECTIVE: Incidental aneurysms pose a challenge for physicians, who need to weigh the rupture risk against the risks associated with treatment and its complications. A statistical model could potentially support such treatment decisions. A recently developed aneurysm rupture probability model performed well in the US data used for model training and in data from two European cohorts for external validation. Because Japanese and Finnish patients are known to have a higher aneurysm rupture risk, the authors' goals in the present study were to evaluate this model using data from Japanese and Finnish patients and to compare it with new models trained with Finnish and Japanese data. METHODS: Patient and image data on 2129 aneurysms in 1472 patients were used. Of these aneurysm cases, 1631 had been collected mainly from US hospitals, 249 from European (other than Finnish) hospitals, 147 from Japanese hospitals, and 102 from Finnish hospitals. Computational fluid dynamics simulations and shape analyses were conducted to quantitatively characterize each aneurysm's shape and hemodynamics. Next, the previously developed model's discrimination was evaluated using the Finnish and Japanese data in terms of the area under the receiver operating characteristic curve (AUC). Models with and without interaction terms between patient population and aneurysm characteristics were trained and evaluated including data from all four cohorts obtained by repeatedly randomly splitting the data into training and test data. RESULTS: The US model's AUC was reduced to 0.70 and 0.72, respectively, in the Finnish and Japanese data compared to 0.82 and 0.86 in the European and US data. When training the model with Japanese and Finnish data, the average AUC increased only slightly for the Finnish sample (to 0.76 ± 0.16) and Finnish and Japanese cases combined (from 0.74 to 0.75 ± 0.14) and decreased for the Japanese data (to 0.66 ± 0.33). In models including interaction terms, the AUC in the Finnish and Japanese data combined increased significantly to 0.83 ± 0.10. CONCLUSIONS: Developing an aneurysm rupture prediction model that applies to Japanese and Finnish aneurysms requires including data from these two cohorts for model training, as well as interaction terms between patient population and the other variables in the model. When including this information, the performance of such a model with Japanese and Finnish data is close to its performance with US or European data. These results suggest that population-specific differences determine how hemodynamics and shape associate with rupture risk in intracranial aneurysms.",
keywords = "AUC = area under the receiver operating characteristic curve, BL = bulge location, cerebral aneurysm, CFD = computational fluid dynamics, hemodynamics, HWR = height/width ratio, IA = intracranial aneurysm, KE = kinetic energy, LSA = low shear area, MLN = mean surface curvature, morphology, NSI = nonsphericity index, OSImax = maximum oscillatory shear stress, risk, rupture, SAH = subarachnoid hemorrhage, WSS = wall shear stress",
author = "Detmer, {Felicitas J.} and Sara Hadad and Chung, {Bong Jae} and Fernando Mut and Martin Slawski and Norman Juchler and Vartan Kurtcuoglu and Sven Hirsch and Philippe Bijlenga and Yuya Uchiyama and Soichiro Fujimura and Makoto Yamamoto and Yuichi Murayama and Hiroyuki Takao and Timo Koivisto and Juhana Fr{\"o}sen and Cebral, {Juan R.}",
year = "2019",
month = "7",
day = "1",
doi = "10.3171/2019.4.FOCUS19145",
language = "English",
volume = "47",
pages = "E16",
journal = "Neurosurgical focus",
issn = "1092-0684",
number = "1",

}

Detmer, FJ, Hadad, S, Chung, BJ, Mut, F, Slawski, M, Juchler, N, Kurtcuoglu, V, Hirsch, S, Bijlenga, P, Uchiyama, Y, Fujimura, S, Yamamoto, M, Murayama, Y, Takao, H, Koivisto, T, Frösen, J & Cebral, JR 2019, 'Extending statistical learning for aneurysm rupture assessment to Finnish and Japanese populations using morphology, hemodynamics, and patient characteristics', Neurosurgical focus, vol. 47, no. 1, pp. E16. https://doi.org/10.3171/2019.4.FOCUS19145

Extending statistical learning for aneurysm rupture assessment to Finnish and Japanese populations using morphology, hemodynamics, and patient characteristics. / Detmer, Felicitas J.; Hadad, Sara; Chung, Bong Jae; Mut, Fernando; Slawski, Martin; Juchler, Norman; Kurtcuoglu, Vartan; Hirsch, Sven; Bijlenga, Philippe; Uchiyama, Yuya; Fujimura, Soichiro; Yamamoto, Makoto; Murayama, Yuichi; Takao, Hiroyuki; Koivisto, Timo; Frösen, Juhana; Cebral, Juan R.

In: Neurosurgical focus, Vol. 47, No. 1, 01.07.2019, p. E16.

Research output: Contribution to journalArticleResearchpeer-review

TY - JOUR

T1 - Extending statistical learning for aneurysm rupture assessment to Finnish and Japanese populations using morphology, hemodynamics, and patient characteristics

AU - Detmer, Felicitas J.

AU - Hadad, Sara

AU - Chung, Bong Jae

AU - Mut, Fernando

AU - Slawski, Martin

AU - Juchler, Norman

AU - Kurtcuoglu, Vartan

AU - Hirsch, Sven

AU - Bijlenga, Philippe

AU - Uchiyama, Yuya

AU - Fujimura, Soichiro

AU - Yamamoto, Makoto

AU - Murayama, Yuichi

AU - Takao, Hiroyuki

AU - Koivisto, Timo

AU - Frösen, Juhana

AU - Cebral, Juan R.

PY - 2019/7/1

Y1 - 2019/7/1

N2 - OBJECTIVE: Incidental aneurysms pose a challenge for physicians, who need to weigh the rupture risk against the risks associated with treatment and its complications. A statistical model could potentially support such treatment decisions. A recently developed aneurysm rupture probability model performed well in the US data used for model training and in data from two European cohorts for external validation. Because Japanese and Finnish patients are known to have a higher aneurysm rupture risk, the authors' goals in the present study were to evaluate this model using data from Japanese and Finnish patients and to compare it with new models trained with Finnish and Japanese data. METHODS: Patient and image data on 2129 aneurysms in 1472 patients were used. Of these aneurysm cases, 1631 had been collected mainly from US hospitals, 249 from European (other than Finnish) hospitals, 147 from Japanese hospitals, and 102 from Finnish hospitals. Computational fluid dynamics simulations and shape analyses were conducted to quantitatively characterize each aneurysm's shape and hemodynamics. Next, the previously developed model's discrimination was evaluated using the Finnish and Japanese data in terms of the area under the receiver operating characteristic curve (AUC). Models with and without interaction terms between patient population and aneurysm characteristics were trained and evaluated including data from all four cohorts obtained by repeatedly randomly splitting the data into training and test data. RESULTS: The US model's AUC was reduced to 0.70 and 0.72, respectively, in the Finnish and Japanese data compared to 0.82 and 0.86 in the European and US data. When training the model with Japanese and Finnish data, the average AUC increased only slightly for the Finnish sample (to 0.76 ± 0.16) and Finnish and Japanese cases combined (from 0.74 to 0.75 ± 0.14) and decreased for the Japanese data (to 0.66 ± 0.33). In models including interaction terms, the AUC in the Finnish and Japanese data combined increased significantly to 0.83 ± 0.10. CONCLUSIONS: Developing an aneurysm rupture prediction model that applies to Japanese and Finnish aneurysms requires including data from these two cohorts for model training, as well as interaction terms between patient population and the other variables in the model. When including this information, the performance of such a model with Japanese and Finnish data is close to its performance with US or European data. These results suggest that population-specific differences determine how hemodynamics and shape associate with rupture risk in intracranial aneurysms.

AB - OBJECTIVE: Incidental aneurysms pose a challenge for physicians, who need to weigh the rupture risk against the risks associated with treatment and its complications. A statistical model could potentially support such treatment decisions. A recently developed aneurysm rupture probability model performed well in the US data used for model training and in data from two European cohorts for external validation. Because Japanese and Finnish patients are known to have a higher aneurysm rupture risk, the authors' goals in the present study were to evaluate this model using data from Japanese and Finnish patients and to compare it with new models trained with Finnish and Japanese data. METHODS: Patient and image data on 2129 aneurysms in 1472 patients were used. Of these aneurysm cases, 1631 had been collected mainly from US hospitals, 249 from European (other than Finnish) hospitals, 147 from Japanese hospitals, and 102 from Finnish hospitals. Computational fluid dynamics simulations and shape analyses were conducted to quantitatively characterize each aneurysm's shape and hemodynamics. Next, the previously developed model's discrimination was evaluated using the Finnish and Japanese data in terms of the area under the receiver operating characteristic curve (AUC). Models with and without interaction terms between patient population and aneurysm characteristics were trained and evaluated including data from all four cohorts obtained by repeatedly randomly splitting the data into training and test data. RESULTS: The US model's AUC was reduced to 0.70 and 0.72, respectively, in the Finnish and Japanese data compared to 0.82 and 0.86 in the European and US data. When training the model with Japanese and Finnish data, the average AUC increased only slightly for the Finnish sample (to 0.76 ± 0.16) and Finnish and Japanese cases combined (from 0.74 to 0.75 ± 0.14) and decreased for the Japanese data (to 0.66 ± 0.33). In models including interaction terms, the AUC in the Finnish and Japanese data combined increased significantly to 0.83 ± 0.10. CONCLUSIONS: Developing an aneurysm rupture prediction model that applies to Japanese and Finnish aneurysms requires including data from these two cohorts for model training, as well as interaction terms between patient population and the other variables in the model. When including this information, the performance of such a model with Japanese and Finnish data is close to its performance with US or European data. These results suggest that population-specific differences determine how hemodynamics and shape associate with rupture risk in intracranial aneurysms.

KW - AUC = area under the receiver operating characteristic curve

KW - BL = bulge location

KW - cerebral aneurysm

KW - CFD = computational fluid dynamics

KW - hemodynamics

KW - HWR = height/width ratio

KW - IA = intracranial aneurysm

KW - KE = kinetic energy

KW - LSA = low shear area

KW - MLN = mean surface curvature

KW - morphology

KW - NSI = nonsphericity index

KW - OSImax = maximum oscillatory shear stress

KW - risk

KW - rupture

KW - SAH = subarachnoid hemorrhage

KW - WSS = wall shear stress

UR - http://www.scopus.com/inward/record.url?scp=85069269235&partnerID=8YFLogxK

U2 - 10.3171/2019.4.FOCUS19145

DO - 10.3171/2019.4.FOCUS19145

M3 - Article

VL - 47

SP - E16

JO - Neurosurgical focus

JF - Neurosurgical focus

SN - 1092-0684

IS - 1

ER -