Intracoronary ultrasound imaging before and after directional coronary atherectomy: in vitro and clinical observations.

The rate of restenosis after directional coronary atherectomy (DCA) is higher than expected. To elucidate why, the current study used intravascular ultrasound (IVUS) imaging to investigate the mechanism of DCA. An in vitro validation study was performed to determine the accuracy of the measurement of plaque removal by IVUS. DCA was performed in eight human atherosclerotic artery segments. The volume of removed plaque was measured by water displacement and was compared with the volume calculated from IVUS images. A clinical study of DCA was performed in 32 lesions. IVUS was performed in 28 lesions after successful DCA. Measurements of lumen dimensions from digital angiograms before and after DCA were compared with observations of lumen and plaque size from the cross-sectional IVUS images. In the in vitro study, the mean plaque volume removed by DCA was 19.9 +/- 8.5 microliters. The calculated estimate of removed plaque volume by IVUS was 18.6 +/- 7.9 microliters and correlated closely with the volume by water displacement (r = 0.92). The calculated volume of plaque removed from histologic sections was 14.3 +/- 6.0 microliters and was linearly correlated with plaque volume by water displacement (r = 0.81). In the clinical study, the angiographic mean minimum lumen diameter increased from 1.0 +/- 0.4 to 2.7 +/- 0.5 mm and the percentage stenosis decreased from 70% to 19% (p < 0.0001). The IVUS images before and after DCA showed that the lumen DCA improved from 2.9 +/- 1.5 to 7.0 +/- 1.5 mm2 (p < 0.0001). In addition the vessel cross-sectional area (CSA) increased from 17.1 +/- 5.9 to 18.7 +/- 5.5 mm2. The atheroma CSA was reduced from 14.2 +/- 5.0 to 11.7 +/- 4.8 mm2. This combined effect of reduction in atheroma CSA and stretching of the outer vessel diameter resulted in an improvement in percentage plaque area stenosis from 83% +/- 7% to 61% +/- 9%. It is concluded that despite a successful angiographic appearance, DCA removed an average of 2.5 mm2 from the atheroma, which corresponds to only 18% of the atheroma CSA. The total lumen CSA increased 4.1 mm2; 61% of the new lumen was created by cutting and removal of plaque, whereas 39% of the new lumen was made by stretching the external wall of the artery. Despite an excellent angiographic result, IVUS imaging reveals that after DCA a significant amount of residual atheroma remains. As in balloon dilatation, a stretching effect is a significant component of DCA.

The rate of restenosis after directional coronary atherectomy (DCA) is higher than expected. To elucidate why, the current study used intravascular ultrasound (IVUS) imaging to investigate the mechanism of DCA. An in vitro validation study was performed to determine the accuracy of the measurement of plaque removal by IVUS. DCA was performed in eight human atherosclerotic artery segments. The volume of removed plaque was measured by water displacement and was compared with the volume calculated from IVUS images. A clinical study of DCA was performed in 32 lesions. IVUS was performed in 28 lesions after successful DCA. Measurements of lumen dimensions from digital angiograms before and after DCA were compared with observations of lumen and plaque size from the cross-sectional IVUS images. In the in vitro study, the mean plaque volume removed by DCA was 19.9 _ 8.5 #1. The calculated estimate of removed plaque volume by IVUS was 18.6 ± 7.9 p.I and correlated closely with the volume by water displacement (r = 0.92). The calculated volume of plaque removed from histologic sections was 14.3 ± 6.0 ~l and was linearly correlated with plaque volume by water displacement (r = 0.81). In the clinical study, the angiographic mean minimum lumen diameter increased from 1.0 ± 0.4 to 2.7 ± 0.5 mm and the percentage stenosis decreased from 70% to 19% (p < 0.0001). The IVUS images before and after DCA showed that the lumen DCA improved from 2.9 ± 1.5 to 7.0 ± 1.5 mm 2 (p < 0.0001). In addition the vessel cross-sectional area (CSA) increased from 17.1 ± 5.9 to 18.7 ± 5.5 mm 2. The atheroma CSA was reduced from 14.2 ± 5.0 to 11.7 ± 4.8 mm 2. This combined effect of reduction in atheroma CSA and stretching of the outer vessel diameter resulted in an improvement in percentage plaque area stenosis from 83% ± 7% to 61% ± 9%. It is concluded that despite a successful angiographic appearance, DCA removed an average of 2.5 mm 2 from the atheroma, which corresponds to only 18% of the atheroma CSA. The total lumen CSA increased 4.1 mm2; 61% of the new lumen was created by cutting and removal of plaque, whereas 39% of the new lumen was made by stretching the external wall of the artery. Despite an excellent angiographic result, IVUS imaging reveals that after DCA a significant amount of residual atheroma remains. As in balloon dilatation, a stretching effect is a significant component of DCA. (AM HEART J 1995;129:841-51.) be obtained compared with standard balloon dilatation. 6-9 However, despite angiographic evidence of a satisfactory acute gain, the rates of restenosis are similar to those after PTCA (27 % to 46%)J°ls The results from the Coronary Angioplasty Versus Excisional Atherectomy Trial (CAVEAT) show a modest improvement in restenosis compared with restenosis after PTCA, but the angiographic restenosis rate for DCA is higher than expected, at 50% .! 9 This observation challenges the assumption that removing atheroma will reduce restenosis; an alternative explanation is that the atherectomy device does not remove as much plaque as predicted. In addition to cutting the plaque, the device may act in a mechanical fashion to stretch the lumen and outer wall of the artery. Angiography is incapable of distinguishing between the cutting effect and the mechanical stretching effect. Intravascular ultrasound (IVUS) imaging provides a method for obtaining quantitative information about the cross-sectional area (CSA) of the lumen and the atheroma burden. 2°-26 By imaging the narrowed segments with IVUS before as well as after atherectomy, it may be possible to obtain a better understanding of the mechanism of action of DCA.

METHODS
In vitro study, To visualize the morphometric effects of DCA and to validate the quantitative findings by IVUS, an in vitro study was performed. Eight human atherosclerotic specimens from the iliac artery were obtained at autopsy. Each artery was cut into 3 cm-long segments. The arteries were stored in normal saline at 4 ° C and were not fixed in formalin until after the images were obtained. The artery segments were connected to an 8F coronary guiding catheter. Two surgical needles were placed through the adventitia of the artery approximately 15 mm apart to provide acoustic reference points. 27 A 25 MHz IVUS catheter (In-terTherapy/CVIS, Sunnyvale, Calif.) was advanced through a Touhy-Borst adaptor on the guiding catheter into the artery segments. The guiding catheter was connected to pressurized normal saline within a closed system to permit imaging under physiologic pressures. The eight arterial segments were imaged before DCA between the two reference needles with use of a motorized pullback device at 0.21 mm/sec. Cross-sectional images were obtained continuously at 30 frames/sec.
After the initial images were obtained, the artery was removed from the guiding catheter platform, and a 7F device (Simpson Atherocath, Devices For Vascular Intervention, Redwood City, Calif.) was placed within the lumen. Multiple passes of the device were made between the two reference needles to acquire material from only one side of the artery. No attempt was made to remove all of the visible atheroma, and one side of the artery was left unaltered to provide a visual comparison of the atherectomy results. The cut pieces of plaque were weighed on an electronic balance. The removed plaque was then dropped into a small graduated cylinder filled with saline with marked gradations of 0.1 ml. The volume of water displaced by the atherectomized plaque was used as the measure of cut plaque volume. After the measurements of weight and volume, the artery was remounted for IVUS imaging after DCA in the same locations relative to the position of the needles in the artery.
At the completion of the study, the arterial segments were pressure perfused for 12 hours in 10% formalin and then placed in decalcification solution for 24 hours. The arteries were imbedded in paraffin after the needle site had been marked with silk suture. Histologic sections were obtained at 1 mm intervals along the length of artery between the reference needle sites. The sections were stained with trichrome. A video image of each histologic section was fed from a microscope (Leica, WILD M3Z, Deerfield, Ill.) through a video camera (Sony, DXC-151A, Tokyo, Japan) into a computer (Macintosh IIci, Apple, Cupertino, Calif.) and digitized (RasterOps 24STV board; Mediagrabber software, Santa Clara, Calif.). Quantitative measurements of the lumen, the plaque, and the atherectomized areas were made with an operator-directed cursor (NIHimage). The IVUS images were digitized from the super-VHS videotape at 0.5 mm intervals along the artery length. The CSAs of the lumen and the plaque before and after atherectomy were measured with the same graphics software as that used for the histologic sections. Each pair of IVUS images (before and after DCA) were placed next to each other to facilitate the identification of the DCA cut. The atherectomized area was measured at each digitized cross-sectional image along the length of the DCA cut. A calculated estimate of plaque volume removed by DCA was obtained by integrating the cut area on each IVUS cross section over the length of the artery segment. A similar calculation of plaque volume was derived from the measurements of the cut areas from the histologic images. The calculated atherectomy plaque volumes from the IVUS images and histologic cross sections were compared by regression analysis with the measured plaque volume by water displacement.

Clinical study
Patient population. Thirty-two lesions in 29 patients were treated with DCA and were studied with IVUS imaging. Of the 32 lesions, 30 (94 %) were treated successfully with DCA alone or adjunctive PTCA. Twenty-four (75 %) lesions had atherectomy performed alone, without adjunctive PTCA. IVUS imaging was performed in 28 lesions after completion of the procedure and was obtained before and after DCA in 19 of the lesions. Patients were selected for DCA rather than conventional PTCA on the basis of presence of localized stenosis in a nontortuous proximal or mid-portion of a major epicardial vessel the reference diameter of which was >--2.5 mm. Successful DCA was defined as removal of tissue, a residual diameter stenosis <50% on digital quantitative angiographic assessment, and absence of major complications (death, myocardial infarction, or emergency coronary artery bypass graft surgery).
DCA. Before initiation of the atherectomy procedure, Lumen CSA was measured at intersection of echolucent lumen and echogenic plaque. Atheroma CSA was then calculated by subtracting lumen CSA from total vessel CSA. Percentage plaque area stenosis was defined as atheroma CSA divided by total available vessel CSA within boundary of media. This definition is similar to histologic definition of percentage area stenosis and differs from angiographic definition of percentage stenosis, which uses proximal segment of vessel as normal reference.
the iliac artery size was determined by angiography to ensure that a 10F or llF sheath would fit adequately in the artery without causing obstruction. The preshaped DCA guiding catheter was positioned at the coronary ostium. Heparin was given to each patient to maintain activated clotting time >300 seconds. Baseline coronary angiograms were obtained in several projections, and then a 0.014inch-diameter flexible guide wire was manipulated across the stenosis. A 3.9F rapid exchange IVUS imaging catheter (InterTherapy/CVIS) was advanced over the guide wire across the stenosis with fluoroscopic control. The catheter consisted of a mechanically rotating 25 MHz ultrasound transducer 1 mm in diameter and was positioned within a 1.3 mm plastic introducing sheath. A full description of the technique of IVUS imaging has been published previously. 24 IVUS imaging was recorded continuously during manual pull-back from the distal segment of the artery to the ostium of the aortic root. Very severe stenoses that would not initially accept either the IVUS catheter or DCA catheter, were predilated with a 2.0 mm balloon. Patients in whom the IVUS catheter could not be passed before DCA were analyzed separately. The DCA catheter was advanced across the stenosis with manual rotation. 2s A 6F or 7F catheter was chosen depending on the size of the artery. The standard recommended procedure was used. 29 Several cuts were made with various orientations of the cutting window as monitored by fluoroscopy. The atherectomy device was withdrawn from the coronary artery after 6 to 10 cuts to remove the tissue in the nosecone. If subsequent angiography showed incomplete resolution of the stenosis, the device was reintroduced and additional cuts were performed using a higher inflation pressure. After DCA was completed as determined by angiography, a second IVUS examination was performed. If the IVUS images showed a significant amount of residual plaque, the information from the IVUS study was used to guide additional DCA cuts.
Analysis of coronary angiograms. Digital coronary angiograms were recorded in at least two projections before and after DCA. These angiograms were stored on computer disk of a digital-acquisition x-ray system (DCI Philips, Skelton, Conn.) and were archived onto 35 mm cine film. After the atherectomy procedure was completed, the digital angiograms were recalled from the x-ray system memory, and the lumen diameter of the treated area and the proximal angiographically normal segment were measured. The diameter of the guiding catheter was used as a reference for calibration. Measurements were made with an operator-defined digital caliper from end-diastolic frames or the frame that best demonstrated the stenotic area. Percentage diameter stenosis was calculated by comparing the diameter of the stenotic segment with an angiographically normal proximal segment as the reference point. Lumen CSA was calculated from orthogonally projected anglograms; internal cross-sectional geometry was assumed to be an ellipse. Lesion length was measured as the distance from the proximal to the distal shoulder of the lesion in the projection that best elongated the stenosis. 3° An eccentric lesion was defined as a stenosis with asymmetry in any angiographic projection. 7 Calcification was defined as a radiopaque density within the vascular wall of the artery at the site of the stenosis. 3° Analysis of IVUS images. IVUS measurements were calculated as demonstrated in Fig. 1. The lumen CSA was defined as the area encompassed by the inner boundary of the intimal surface. The major and minor lumen diameters were measured at the long axis and short axis of the lumen. The vessel CSA was defined as the area bounded by the hypoechoic medial ring. If the vessel area could not be traced because of calcium, a proximal adjacent noncalcified site was measured as the vessel area. The major and minor vessel diameters were measured at the long axis and short axis of the vessel at the border between the atheroma and media. The atheroma CSA was calculated as the vessel CSA   minus the lumen CSA. The percentage area stenosis was defined as the atheroma CSA divided by the vessel CSA. This approach is similar to a histologic definition of stenosis and does not depend on a proximal reference segment. All measurements were performed during diastole.
To quantitate plaque eccentricity, an eccentricity index was calculated from the IVUS image as the thinnest dimension of the plaque divided by the width of the opposite wall. 25 A stenosis was defined as concentric if the atherosclerotic plaque was distributed circumferentially, corresponding to an eccentricity index _>0.50. The stenosis was defined as eccentric if the index was <0.50. Calcified tissue was defined as an area of intense echogenicity with acoustic shadowing. The calcium score was quantified as absent (0), mild (1+) when calcification was localized to _<90 degrees of arc, moderate (2+) when it involved 91 to 180 degrees of arc, severe (3+) when it incorporated 181 to 270 degrees of arc, and (4+) when >270 degrees of arc was calcified.
Statistics. Group data were tabulated as means + SDI Comparison of the CSA of the lumen, vessel, plaque, and percentage stenosis before and after the atherectomy procedure were performed by the paired Student's t test. Two-by-two comparison tables were assessed by chi square analysis. Differences were considered to be statistically significant at p < 0.05. Linear regression analysis was performed to compare different measurements. Comparisons between more than two groups were performed by analysis of variance. The variability of IVUS measurements from our laboratory has been reported previously. 31 Correlation coefficients for intra-and interobserver measurements of minimum lumen diameter (MLD) and lumen CSA ranged from 0.92 to 0.99 after PTCA.

In vitro study.
T h e m e a n plaque volume removed by the DCA device from the eight artery segments was 19.9 +_ 8.5 #l as measured by water displacement. T h e m e a n weight was 21.0 + 7.0 mg. Representative cross-sectional I V U S images before and after DCA c o m p a r e d with the histologic section are presented in Fig. 2. All of the artery segments d e m o n s t r a t e d the typical groove cut into the fibrous plaque capsule. T h e DCA plaque volume calculated from the ultrasound images was 18.6 + 7.9 #l, and the value calculated from the histologic cross sections was 14.3 _+ 6.0 /~1. T h e correlation between the m e a s u r e m e n t s of cut plaque volume between IVUS images and histologic sections was r = 0.74. T h e correlation coefficients between the measured plaque volume by water disp l a c e m e n t compared with histologic sections and IVUS images are presented in T a b l e I. The comparison of the IVUS images before and after DCA showed significant variability in lumen dimensions caused by stretching of the artery during the manipulation to mechanically cut the plaque from the arterial segments. The mean lumen CSA before atherectomy was 25.4 + 2.2 mm 2 and was 28.2 _+ 1.9 mm 2 after DCA (p < 0.05).
Clinical study. Table II summarizes the clinical characteristics of the patients who were treated with DCA during this study. Table III lists the stenosis and procedure characteristics. The average lesion length was 10.1 + 4.8 mm. Only 4 of the stenoses showed obvious calcification by cineangiographic criteria. Eighteen (54 % ) of the stenoses were eccentric by angiography. The number of cuts obtained with the atherectomy catheter varied between 7 and 73 (average of 25 + 22 cuts). The average balloon inflation pressure during atherectomy was 28 + 16 (range 10 to 60) psi. Twenty-four (75 % ) lesions were successfully treated with DCA alone. Because of the severity of disease, PTCA before DCA was performed in six patients to facilitate passage of the atherectomy catheter. Six (19%) patients also received adjunctive PTCA after DCA.
In two patients DCA was unsuccessful. In 1 case, the procedure was performed during an acute myocardial infarction in the proximal right coronary artery. It was difficult to deliver the DCA device because of superior angulation in the proximal segment. The other patient had previous bypass surgery with severe disease in the left main coronary artery. There was circumferential dense calcification in the area of stenosis as revealed by IVUS examination before DCA. It was difficult to establish proper positioning of the guiding catheter in this case, and the  DCA device could not be delivered appropriately. These two patients underwent successful PTCA alone.
Quantitative results. The angiographic measurements before and after DCA or DCA plus PTCA are listed in Table IV. The mean MLD increased significantly, from 1.0 + 0.4 mm to 2.7 _+ 0.5 mm, and the percentage stenosis decreased from 70% to 19% (p < 0.0001). There was a smooth angiographic appearance in 23 lesions; 5 lesions had a hazy luminal border; and 2 revealed a dissection (1 resulting from adjunctive PTCA).
Table V provides the quantitative measurements from the IVUS studies before and after DCA at the same cross section. Data are presented only for the  patients who received DCA without adjunctive P T C A to prevent mixing the effects of these two procedures (n = 19). After DCA, the lumen CSA improved significantly, from 2.9 _+ 1.5 mm 2 to 7.0 +_ 1.5 mm 2.
Major and minor lumen diameters also were improved, by 70 % and 47 9 , respectively. In addition, the vessel CSA increased from 17.1 +_ 5.9 mm 2 to 18.7 + 5.5 mm 2. This increase corresponded to an increase in the diameter of the vessel (measured at the boundary of the media and atheroma) from stretching. T h e atheroma CSA was reduced from 14.2 + 5.0 mm 2 to 11.7 _+ 4.8 m m 2. This combined effect of reduction in atheroma CSA and stretching of the outer vessel diameter resulted in an improvement in percentage area stenosis from 83 % _+ 7 % to 61% + 9% (Fig. 3). The calcium score as measured from the IVUS images was 1.5 + 0.7 before and 1.2 _+ 0.9 after DCA (p not significant). This indicates t h a t the vessels chosen for DCA were not heavily calcified nor was there a significant amount of calcium removed from the plaque (Fig. 4). T h e majority of these lesions were mildly concentric stenoses as measured by intravascular cross-sectional imaging, with a mean eccentricity index of 0.57 _+ 0.27 at baseline. This index was reduced to 0.35 + 0.23 after DCA. T h e increase in the eccentricity of the plaques indicates t h a t the atherectomy cuts were not circumferentially uniform but tended to remain in one quadrant despite attempts to rotate the cutter (Fig. 5).
Figs. 6 and 7 demonstrate the correlation between IVUS and angiography for the M L D and lumen CSA after DCA without adjunctive PTCA. T h e tightest angiographic M L D was selected from at least two   5. This 83-year-old man had class I angina with tandem lesions in proximal and mid-right coronary artery (A). There was no evidence of calcium by fluoroscopy. IVUS revealed 3+ calcification in proximal lesion (B) and 2+ calcification in mid lesion (C) in this vessel. After multiple cuts, IVUS was performed but revealed significant residual atheroma in the vessel wall. Additional DCA cuts were performed. In the mid lesion, a small amount of additional plaque was removed (12 o'clock) (F). In proximal lesion (E), although atheroma area did not change, lumen was stretched, resulting in satisfactory angiographic result (D).  projections. If the lesion was observed in only one projection, the lesion was excluded for analysis. Fifteen lesions were obtained with orthogonal projections. There was a significant correlation of MLD between angiography and IVUS after DCA (r = 0.73; p = 0.002). There also was a significant correlation of lumen CSA between both measurements after DCA (r = 0.74; p = 0.0015).

Correlation of lumen CSA between IVUS and an-
Complications. There were no instances of myocardial infarction, emergency bypass surgery, or death during any of these procedures. One patient had ventricular tachycardia during DCA associated with prolonged ischemia from the guiding catheter and the cutter. The patient recovered after removal of the catheter and electric cardioversion. The procedure was continued and yielded a successful result with significant improvement in blood flow.

DISCUSSION
The IVUS observations obtained in this study help to explain the mechanism of DCA more accurately than does angiography. DCA removed an average of 2.5 mm 2 from the atheroma, which corresponds to 18% of the atheroma CSA. The total lumen CSA increased 4.1 mm2; thus 61% of the new lumen was created by DCA, and 39 % of the new lumen was made by stretching the external wall of the artery (Fig. 8).
It has been assumed that the fundamental mechanism of DCA is the excision of plaque. However, lumen diameter improvement on angiography does not prove that all of the effect is attributable to removal of plaque. Safian et al. 32 concluded that lumen enlargement with DCA is caused in part by mechanical dilatation in addition to tissue removal. Their conclusion was derived from angiographic observations compared with the amount of tissue removed. In another angiographic study Penny et al. 33 concluded that the average cutting effect contributed 28 % of the new lumen. Our study corroborates this data by direct observation with IVUS imaging. In contrast to our findings, Tenaglia et al. 34 using IVUS reported that the effect of stretching during DCA was small.
In vitro study. The results from the companion in vitro studies demonstrate that IVUS imaging accurately represents the amount of plaque that is removed by DCA. Compared with direct measurement by water displacement of removed plaque volume, IVUS provided a close estimate of the mean plaque volume (19.9 vs 18.6 ul). In addition the individual IVUS measurements correlated more closely with water displacement (r = 0.92) than did histologic measurements (r = 0.81). These observations give credibility to the IVUS measurements obtained during the clinical studies. During the clinical cases, measurements were made of the residual lumen and plaque CSAs. Although plaque volume calculations were not performed during the clinical studies, the method of integrating the areas over the length of the treated artery could be applied at the time of the procedure to quantitate the volume of plaque removed. The in vitro IVUS images showed that a successful DCA demonstrated a smooth groove or "bite" in the atheroma, without separation of plaque from the media. In distinction to a previous study using this imaging technology to understand the mechanism of PTCA no intimal dissections were observed with the DCA device. 27 Clinical study. The angiographic results in our patients are similar to the results from larger series of patients that were analyzed with angiography only. Popma et al. 35 reported that the lumen CSA increased from 1.2 + 0.9 mm 2 to 6.4 + 4.4 mm 2 after DCA in left anterior descending lesions. 35 Ellis et al. 7 demonstrated that the lumen dimensions increased from 1.1 +_ 1.0 mm to 2.6 + 2.3 mm. The diameter stenosis reported by Mueller et al. 36 improved from 69% _+ 10% to 22% + 20%. These studies did not use IVUS imaging to assess the effect of DCA.
Angiography versus IVUS. There was a significant correlation between angiography and IVUS for the measurements of lumen diameter after DCA (r = 0.73). This is in distinction to previous studies using IVUS to assess the results of PTCA alone. 24 In those series the correlation coefficient range was r = 0.12 to 0.21 after PTCA. It was concluded that the poor correlation was due to the dissections produced by balloon angioplasty. Contrast streaming into these dissection planes produces a wider apparent diameter on projection imaging techniques compared with that produced by cross-sectional imaging methods such as IVUS. Garratt et al. 37 showed no plaque fracturing or medial dissection after DCA in an autopsy study. The current study also demonstrated a low rate of dissection following DCA in vivo or dur-ing the in vitro studies. Penny et al. 33 suggested that DCA produces cloverlike bites into the plaque. Although this may be true for larger peripheral arteries, in coronary arteries the cross-sectional IVUS images suggest that an eccentric single trough is the usual morphologic result.
The majority of the stenoses that were studied before DCA had a concentric morphology by IVUS. The IVUS description of eccentricity corresponds to a histologic definition using measurements of the maximum and minimum thickness of the plaque. A plaque wall ratio >0.5 is considered a concentric stenosis with relatively equal distribution of plaque around the circumference of the lumen, whereas a major to minor plaque wall thickness of <0.5 is defined as an eccentric stenosis. The mean eccentricity index was 0.57 _+ 0.27 at baseline, consistent with a definition of mildly concentric stenoses. After DCA the lesions were transformed to a predominately eccentric morphology with an index of 0.35 + 0.23 (p < 0.01). This observation indicates that DCA did not cut the atheroma uniformly despite the attempt to rotate the atherectomy canoe radially and the use of IVUS to redirect the cutting assembly. This result may be caused by areas along the circumference of the plaque that are more resistant to the cutting effect of the DCA device (perhaps because of dense fibrosis or calcium). Alternatively, it is possible that once an initial trough is made in the atheroma, the canoe tends to slip back into the same position, resulting in an eccentric morphology. This may help explain the observation that despite an increased number of cutting passes, no more material may be removed. Whether a plaque is designated to be concentric or eccentric may have a significant effect on the treatment strategy. This is usually defined angiographically, and the information is used by the operator to direct the cutting blade. However, the results of IVUS imaging demonstrate a poor correlation between angiography and IVUS for the diagnosis of eccentricity. 25 In the present study, 55% of angiographic eccentric lesions were found to be concentric on cross-sectional IVUS imaging. In addition, 70 % of angiographic concentric lesions were found by IVUS to be eccentric. This information was used by the operator in an iterative process to alter the approach and direction of the DCA device. In addition, the information provided by the IVUS images was used to perform further DCA cuts despite an adequate angiographic result in 17 % of cases.
The low calcium score indicates a bias of patient selection for less calcified lesions. The calcium score did not improve after DCA, consistent with previous observations that DCA is less effective for calcified lesions 3s especially if the calcium is located superficially at the lumen interface. 39, 40 Cross sectional imaging by IVUS is more explicit than angiography in defining the presence and location of calcified plaque (Fig. 9).
Conclusions. Although DCA produces significant improvement in angiographic lumen dimensions, ob-servations from IVUS imaging reveal that 40 % of the angiographic effect is due to stretching of the external vessel wall. There is frequently a large residual plaque after DCA which encompasses a mean of 59 % of the available vessel CSA. These findings may help to explain the unexpectedly high incidence of restenosis after DCA. Because only 18 % of the atheroma CSA is removed by DCA, it would require a relatively minor amount of cellular matrix proliferation and elastic recoil to produce restenosis with a 50 % lumen diameter.