Unsupervised image classification of medical ultrasound data by multiresolution elastic registration.

Thousands of medical images are saved in databases every day and the need for algorithms able to handle such data in an unsupervised manner is steadily increasing. The classification of ultrasound images is an outstandingly difficult task, due to the high noise level of these images. We present a detailed description of an algorithm based on multiscale elastic registration capable of unsupervised, landmark-free classification of cardiac ultrasound images into their respective views (apical four chamber, two chamber, parasternal long axis and short axis views). We validated the algorithm with 90 unselected, consecutive echocardiographic images recorded during daily clinical work. When the two visually very similar apical views (four chamber and two chamber) are combined into one class, we obtained a 93.0% correct classification (chi2 = 123.8, p < 0.0001, cross-validation 93.0%; chi2 = 131.1, p < 0.0001). Classification into the 4 classes reached a 90.0% correct classification (chi2 = 205.4, p < 0.0001, cross-validation 82.2%; chi2 = 165.9, p < 0.0001).

Patterns of changes in arterial PO2 during one-lung ventilation: a comparison between patients with severe pulmonary emphysema and patients with preserved lung function.

OBJECTIVES: One-lung ventilation (OLV) during thoracoscopic surgery is associated with a significant decline in arterial PO(2) in patients with severe pulmonary emphysema and patients with preserved lung function. The authors hypothesized that patterns of arterial PO(2) changes are different in these 2 patient groups. DESIGN: Prospective nonrandomized study. SETTING: University hospital. PARTICIPANTS: Twenty-five patients undergoing thoracoscopic interventions: 16 with severe pulmonary emphysema and 9 patients without emphysema. INTERVENTIONS: Continuous arterial blood gas measurement (PaO(2), PaCO(2), pHa) during OLV of the left lung in left lateral position using the Paratrend 7 blood gas monitoring system (PT7; Pfizer Hospital Products Group, High Wycombe, UK). MAIN RESULTS: The decrease of PaO(2) was delayed in patients with severe emphysema. Steady state (defined as DeltaPaO(2) <7.5 mmHg/min) was reached after 18 +/- 4 minutes compared with 11 +/- 3 minutes (mean +/- standard deviation) in patients with normal lung function (p = 0.0002). PaO(2) values at steady state were comparable (p = 0.49); the pattern of changes in PaO(2) for the first 15 minutes of left-sided OLV was significantly different between the groups (p = 0.0004). The difference of predicted versus measured PaO(2) at steady state was -48 +/- 160 mmHg for patients with emphysema and -51 +/- 60 mmHg for patients with normal lung function (p = 0.019). CONCLUSION: During OLV, oxygenation is better preserved for a longer period of time in patients with severe pulmonary emphysema as compared with patients with normal lung function. In contrast to patients without emphysema, prediction of oxygenation during OLV for the individual patient with emphysema is unreliable because of large interindividual differences.

Novel doppler assessment of intracoronary volumetric flow reserve: validation against PET in patients with or without flow-dependent vasodilation.

UNLABELLED: Volumetric blood flow (Q) determination requires simultaneous assessment of mean blood flow velocity and vessel cross-sectional area. At present, no method provides both values. Intracoronary Doppler-based assessment of coronary flow velocity reserve (CFVR) relies on average peak velocity (APV). Because this does not account for changes in velocity profile or vessel area usually occurring with flow-dependent vasodilation, results can be misleading. The aim of this clinical study was to validate against the current gold standard (measurement of myocardial perfusion reserve [MPR] by PET) a new, Doppler-based method for calculating coronary Q and coronary flow reserve (CFR). METHODS: Doppler-based intracoronary Q was measured with a proprietary guidewire device in a nonstenotic coronary artery at baseline and during adenosine-induced hyperemic flow (140 mug/kg/min intravenously during 7 min). Three gate positions were assessed, of which 2 were lying within the vessel and 1 was intersecting the vessel. The zeroth (M(0)) and the first (M(1)) Doppler moments of the intersecting gate were used to calculate mean blood flow velocity (M(1)/M(0)) and vessel area (M(0)), and M(0) of the 2 proximal gates was used to correct for scattering and attenuation. CFR was calculated as hyperemic/resting flow with Q and compared with APV-derived CFVR and with the corresponding segmental MPR obtained with (15)O-labeled water and PET. RESULTS: Q (CFR, 2.60 +/- 1.07) correlated well with PET (MPR, 2.58 +/- 1.11) (r = 0.832, P < 0.005; Bland-Altman limits, -1.42 to 1.09), whereas CFVR did not (r = 0.09, P = not statistically significant; Bland-Altman limits, -3.36 to 2.24). However, in vessels without dilation, there was no difference between CFR, CFVR, and MPR. CONCLUSION: This procedure for intracoronary Q measurement using the proprietary Doppler guidewire system, which accounts for both changes in flow profile and changes in vessel area, allows invasive, accurate assessment of CFR even in the presence of flow-dependent vasodilation.

Texture analysis in digitally-acquired echocardiographic images: the effect of JPEG compression and video storage.

The analysis of texture in video-stored echocardiographic images is an established method to characterize myocardial pathologies. We investigated whether or not texture parameters calculated from video-stored images and those derived from the joint photographic expert group (JPEG) format compressed data are equivalent to those calculated from uncompressed digital images. Texture parameters were calculated using uncompressed digital data, images stored on videotape, and three forms of compressed digital data (baseline JPEG, JPEG 2000 and lossless JPEG 2000). Video storage heavily affected most texture parameters. Although first-order texture parameters derived from JPEG-compressed images were generally equivalent to those derived from the uncompressed data, several second-order parameters differed significantly. We conclude that texture of video-stored images is not comparable to that of digitally-stored images and that JPEG compression changes important second-order texture parameters. This observation should be taken into account when analyzing texture of modern image data (uncompressed or compressed) and comparing the results with earlier studies utilizing video-stored data.

A novel in vivo procedure for volumetric flow measurements.

We report on a novel procedure for invasive volumetric blood flow measurements using a commercially available Doppler flow wire system, which could, until now, only measure flow velocity. We here describe a method applicable in vivo to generate both velocity and cross-sectional area information from the same pulsed-wave Doppler signal for volumetric flow assessment. We demonstrate its feasibility and validation in vivo in pig coronary arteries. Our Doppler-derived volumetric flow measurements were compared with the respective transit-time flow and showed an excellent correlation (r = 0.969; p < 0.0001). Agreement between transit-time and Doppler-derived flow measurements could be observed for flow conditions ranging from 30 to 180 mL/min. The mean values for the two methods were 71.4 +/- 43.7 mL/min and 71.3 +/- 42.2 mL/min, respectively. We conclude that this technique might possibly be introduced into future clinical practice as an invasive procedure of choice for the assessment of volumetric blood flow.