Ultraschall Med 2013; 34(2): 145-150
DOI: 10.1055/s-0031-1299331
Original Article
© Georg Thieme Verlag KG Stuttgart · New York

Automatic Texture-Based Analysis in Ultrasound Imaging of Ovarian Masses

Automatische Texturanalysen in der Ultraschallbildgebung von Ovarialbefunden
F. Faschingbauer
1   Gynäkologie und Geburtshilfe, Universitätsfrauenklinik Erlangen
,
M. W. Beckmann
1   Gynäkologie und Geburtshilfe, Universitätsfrauenklinik Erlangen
,
T. Weyert Goecke
1   Gynäkologie und Geburtshilfe, Universitätsfrauenklinik Erlangen
,
S. Renner
1   Gynäkologie und Geburtshilfe, Universitätsfrauenklinik Erlangen
,
L. Häberle
1   Gynäkologie und Geburtshilfe, Universitätsfrauenklinik Erlangen
,
M. Benz
2   Institute for Integrated Circuits (IIS), Frauenhofer Institute
,
T. Wittenberg
2   Institute for Integrated Circuits (IIS), Frauenhofer Institute
,
C. Münzenmayer
2   Institute for Integrated Circuits (IIS), Frauenhofer Institute
› Author Affiliations
Further Information

Publication History

11 November 2011

13 December 2011

Publication Date:
23 May 2012 (online)

Abstract

Purpose: To assess the diagnostic accuracy of a new automatic texture-based algorithm (ATBA) in ultrasound imaging of ovarian masses and to compare its performance to subjective assessment by examiners with different levels of ultrasound experience.

Materials and Methods: A total of 105 ultrasound images from three different groups of ovarian lesions (malignancies, functional cysts, and dermoid cysts) were evaluated using ATBA and by a total of 36 examiners with four different levels of experience (9 junior trainees, 8 senior trainees, 11 senior gynecologists, and 8 experts). Cohen’s κ, Youden’s indices, and the sensitivity and specificity of ATBA and of each observer were calculated for every subgroup of ovarian lesions.

Results: ATBA classified 78 of the 105 masses correctly (κ = 0.62) – results that were significantly better than those of the junior and senior trainees (p = 0.02 and p < 0.01), while differences from the group of level II examiners did not reach statistical significance (p = 0.27). The best diagnostic performance (κ = 0.70) was obtained by the group of expert level III ultrasonographers. The best classification rates overall, including both ATBA and subjective assessments, were achieved in the detection of functional cysts (Youden’s indices from 0.73 to 0.85), while the poorest diagnostic performance was obtained for the classification of dermoid cysts (Youden’s indices from 0.28 to 0.55).

Conclusion: ATBA showed a significantly better diagnostic performance than observers with low or medium levels of experience, emphasizing its potential value for training purposes and in providing additional diagnostic assistance for inexperienced observers.

Zusammenfassung

Ziel: Evaluierung automatischer Texturanalysen (ATA) in der Ultraschallbildgebung von Ovarialbefunden und Vergleich mit der subjektiven Einschätzung von Untersuchern mit unterschiedlichen Graden an klinischer Erfahrung.

Material und Methoden: Insgesamt 105 Ultraschallbilder von malignen Ovarialtumoren, funktionellen Zysten und Dermoid-Zysten wurden sowohl durch ATA als auch subjektiv von 36 Untersuchern mit 4 verschiedenen Graden an klinischer Erfahrung analysiert. Die Untersuchergruppen bestanden aus 9 unerfahrenen sowie 8 erfahreneren Assistenzärzten (Level-I-Untersucher), 11, nicht auf den Ultraschall spezialisierten Oberärzten (Level-II-Untersucher) und 8 Experten (Level-III-Untersucher) mit Schwerpunkt im Bereich des gynäkologischen Ultraschalls. Cohens κ, Youdens-Indices, Sensitivität und Spezifität wurden sowohl für die ATA als auch für jeden Untersucher berechnet.

Ergebnisse: ATA erzielte signifikant bessere Ergebnisse als die beiden Gruppen von Assistenzärzten (κ = 0,62; p = 0,02 im Vergleich zu den unerfahrenen Untersuchern und p < 0,01 für Level-I-Untersucher). Die Unterschiede zu den Level-II-Untersuchern erreichte keine statistische Signifikanz (p = 0,27). Die signifikant besten Ergebnisse wurden von den Level-III-Untersuchern erzielt (κ = 0,70; p < 0,01 im Vergleich zu ATA). Für die Klassifikation von funktionellen Zysten (Youdens-Indices von 0,73 bis 0,85), ergaben sich sowohl mit ATA als auch in der subjektiven Beurteilung die besten Werte, wohingegen die schlechtesten Ergebnisse für die Detektion von Dermoid-Zysten (Youdens-Indices von 0,28 bis 0,55) erreicht wurden.

Schlussfolgerung: ATA erreichte signifikant bessere Ergebnisse als Untersucher mit einem niedrigen und mittleren Grad an klinischer Erfahrung und zeigt damit sein Potenzial im Rahmen der klinischen Ausbildung und als diagnostische Hilfestellung für unerfahrene Untersucher.

 
  • References

  • 1 Borgfeldt C, Andolf E. Transvaginal sonographic ovarian findings in a random sample of women 25-40 years old. Ultrasound Obstet Gynecol 1999; 13: 345-350
  • 2 Sankaranarayanan R, Ferlay J. Worldwide burden of gynaecological cancer: the size of the problem. Best Pract Res Clin Obstet Gynaecol 2006; 20: 207-225
  • 3 Killackey MA, Neuwirth RS. Evaluation and management of the pelvic mass: a review of 540 cases. Obstet Gynecol 1988; 71: 319-322
  • 4 Eriksson L, Kjellgren O, von Schoultz B. Functional cyst or ovarian cancer: histopathological findings during 1 year of surgery. Gynecol Obstet Invest 1985; 19: 155-159
  • 5 Caspi B, Appelman Z, Rabinerson D et al. The growth pattern of ovarian dermoid cysts: a prospective study in premenopausal and postmenopausal women. Fertil Steril 1997; 68: 501-505
  • 6 Valentin L. Imaging in gynecology. Best Pract Res Clin Obstet Gynaecol 2006; 20: 881-906
  • 7 Merz E, Weber G, Bahlmann F et al. A new sonomorphologic scoring system (Mainz Score) for the assessment of ovarian tumors using transvaginal ultrasonography. Part I: A comparison between the scoring-system and the assessment by an experienced sonographer. Ultraschall in Med 1998; 19: 99-107
  • 8 Weber G, Merz E, Bahlmann F et al. A new sonomorphologic scoring-system (Mainz score) for the assessment of ovarian tumors using transvaginal ultrasonography. Part II: A comparison between the scoring-system and the assessment by an experienced sonographer in postmenopausal women. Ultraschall in Med 1999; 20: 2-8
  • 9 Valentin L. Use of morphology to characterize and manage common adnexal masses. Best Pract Res Clin Obstet Gynaecol 2004; 18: 71-89
  • 10 Valentin L, Ameye L, Testa A et al. Ultrasound characteristics of different types of adnexal malignancies. Gynecol Oncol 2006; 102: 41-48
  • 11 Yazbek J, Raju KS, Ben-Nagi J et al. Accuracy of ultrasound subjective “pattern recognition” for the diagnosis of borderline ovarian tumors. Ultrasound Obstet Gynecol 2007; 29: 489-495
  • 12 Muller F, Stiepani H, Unger M et al. Ultrasonographic Appearance of a Dermoid Cyst Completely Lined with Colonic Mucosa. Geburtshilfe Und Frauenheilkunde 2010; 70: 825-827
  • 13 Valentin L. Prospective cross-validation of Doppler ultrasound examination and gray-scale ultrasound imaging for discrimination of benign and malignant pelvic masses. Ultrasound Obstet Gynecol 1999; 14: 273-283
  • 14 Van Calster B, Timmerman D, Bourne T et al. Discrimination between benign and malignant adnexal masses by specialist ultrasound examination versus serum CA-125. J Natl Cancer Inst 2007; 99: 1706-1714
  • 15 Valentin L. Pattern recognition of pelvic masses by gray-scale ultrasound imaging: the contribution of Doppler ultrasound. Ultrasound Obstet Gynecol 1999; 14: 338-347
  • 16 Timmerman D, Schwarzler P, Collins WP et al. Subjective assessment of adnexal masses with the use of ultrasonography: an analysis of interobserver variability and experience. Ultrasound Obstet Gynecol 1999; 13: 11-16
  • 17 Schulz-Wendtland R, Wenkel E, Wacker T et al. Quo vadis? Trends in Digital Mammography. Geburtshilfe Und Frauenheilkunde 2009; 69: 108-117
  • 18 Elter M, Schulz-Wendtland R, Wittenberg T. The prediction of breast cancer biopsy outcomes using two CAD approaches that both emphasize an intelligible decision process. Med Phys 2007; 34: 4164-4172
  • 19 Munzenmayer C, Kage A, Wittenberg T et al. Computer-assisted diagnosis for precancerous lesions in the esophagus. Methods Inf Med 2009; 48: 324-330
  • 20 Grab D, Merz E, Prompeler H et al. Standards for ultrasound in gynecology. Ultraschall in Med 2011; 32: 415-417
  • 21 Gembruch U, Merz E. Standardization of ultrasound diagnostics in gynecology. Ultraschall in Med 2011; 32: 339-341
  • 22 EFSUMB. Minimum training recommendations for the practice of medical ultrasound. Ultraschall in Med 2006; 27: 79-105
  • 23 Chen YQ NM, Thomas DW. Statistical geometrical features for texture classification. Pattern Recognit 1995; 28: 537-552
  • 24 Burges C. A Tutorial on Support Vector Machines for Pattern Recognition. Data Mining and Knowledge Discovery 1998; 2: 121-167
  • 25 Fan RE, Chen PH, Lin CJ. Working set selection using the second order information for training SVM. Journal of Machine Learning Research 2005; 6: 1889-1918
  • 26 Chang CC, Lin CJ. LIBSVM: a library for support vector machines. ACM Transactions on Intelligent Systems and Technology 2011; 2: 1-27
  • 27 Zimmer Y, Tepper R, Akselrod S. An automatic approach for morphological analysis and malignancy evaluation of ovarian masses using B-scans. Ultrasound Med Biol 2003; 29: 1561-1570
  • 28 Patel MD, Feldstein VA, Chen DC et al. Endometriomas: diagnostic performance of US. Radiology 1999; 210: 739-745
  • 29 Alcazar JL, Guerriero S, Laparte C et al. Diagnostic performance of transvaginal gray-scale ultrasound for specific diagnosis of benign ovarian cysts in relation to menopausal status. Maturitas 2011; 68: 182-188
  • 30 Van Holsbeke C, Daemen A, Yazbek J et al. Ultrasound experience substantially impacts on diagnostic performance and confidence when adnexal masses are classified using pattern recognition. Gynecol Obstet Invest 2010; 69: 160-168
  • 31 Guerriero S, Alcazar JL, Pascual MA et al. Diagnosis of the most frequent benign ovarian cysts: is ultrasonography accurate and reproducible?. J Womens Health (Larchmt) 2009; 18: 519-527
  • 32 Van Holsbeke C, Yazbek J, Holland TK et al. Real-time ultrasound vs. evaluation of static images in the preoperative assessment of adnexal masses. Ultrasound Obstet Gynecol 2008; 32: 828-831
  • 33 Stein SM, Laifer-Narin S, Johnson MB et al. Differentiation of benign and malignant adnexal masses: relative value of gray-scale, color Doppler, and spectral Doppler sonography. Am J Roentgenol 1995; 164: 381-386