Importance of the choice of the collimator for the detection of small lesions in scintimammography: a phantom study

被引:13
作者
Buvat, I
Laffont, S
Le Cloirec, J
Bourguet, P
Di Paola, R
机构
[1] CHU Pitie Salpetriere, INSERM U494, F-75634 Paris 13, France
[2] Ctr Reg Lutte Contre Canc Eugene Marquis, Nucl Med Serv, F-35062 Rennes, France
关键词
D O I
10.1088/0031-9155/46/5/301
中图分类号
R318 [生物医学工程];
学科分类号
0831 ;
摘要
Tc-99m methoxyisobutylisonitrile planar scintimammography (SMM) is mostly performed using low-energy high-resolution (LEHR) parallel collimators. We studied whether using a different collimator could improve the detection of small (<1.5 cm) lesions for which SMM sensitivity is poor. Thirty four breast phantom configurations were considered, either with hot spheres simulating lesions or without any spheres. For each configuration, four planar acquisitions were performed using LEHR, low-energy ultra high-resolution (LEUHR), high-resolution fan-beam (HRFB) and ultra high-resolution fan-beam (UHRFB) collimators. Images corresponding to the 20% and 10% energy windows and to the Jaszczak subtraction were calculated. A database including 156 borderline images was derived. After training, 10 observers scored the images for the presence of a sphere. The performances in sphere detection were studied using receiver operating characteristic (ROC) analysis. For all types of image, the area under the ROC curve was highest with the UHRFB collimator and lowest with either the LEUHR or the HRFB collimator. For the 10% energy window images conventionally used in SMM, the detection sensitivities averaged 91%, 73%, 60% and 55% for the UHRFB, LEHR, HRFB and LEUHR collimators respectively, for the same specificity of 64%. We conclude that detection of small tumours in planar SMM might be significantly improved by using a UHRFB collimator instead of an LEHR collimator.
引用
收藏
页码:1343 / 1355
页数:13
相关论文
共 17 条
[1]  
Buvat I, 1998, J NUCL MED, V39, P1590
[2]   MAXIMUM-LIKELIHOOD ESTIMATION OF PARAMETERS OF SIGNAL-DETECTION THEORY AND DETERMINATION OF CONFIDENCE INTERVALS - RATING-METHOD DATA [J].
DORFMAN, DD ;
ALF, E .
JOURNAL OF MATHEMATICAL PSYCHOLOGY, 1969, 6 (03) :487-&
[3]   THE MEANING AND USE OF THE AREA UNDER A RECEIVER OPERATING CHARACTERISTIC (ROC) CURVE [J].
HANLEY, JA ;
MCNEIL, BJ .
RADIOLOGY, 1982, 143 (01) :29-36
[4]   SCATTER COMPENSATION TECHNIQUES FOR SPECT [J].
JASZCZAK, RJ ;
FLOYD, CE ;
COLEMAN, RE .
IEEE TRANSACTIONS ON NUCLEAR SCIENCE, 1985, 32 (01) :786-793
[5]  
KHALKHALI I, 1994, J AM COLL SURGEONS, V178, P491
[6]  
Khalkhali I, 1999, J NUCL MED, V40, P1233
[7]  
Maini CL, 1999, J NUCL MED, V40, P46
[8]  
Maublant J, 1996, J NUCL MED, V37, P922
[9]  
Mekhmandarov S, 1998, J NUCL MED, V39, P86
[10]   BASIC PRINCIPLES OF ROC ANALYSIS [J].
METZ, CE .
SEMINARS IN NUCLEAR MEDICINE, 1978, 8 (04) :283-298