DE102019111617B4 - Process for assessing / determining / calculating the cardiovascular risk profile of a person by evaluating a chest CT of that person - Google Patents

Abstract

A method for determining / evaluating / assessing the cardiovascular risk profile of a person by evaluating a thorax computed tomography of this person is characterized in that to determine the total length of a calcification - each thorax CT slice of the vertically running portions of the LM, LAD , LCX and RCA coronary arteries, in which a point-like calcification ≤ 2mm is identified, counts, and - counts the number of layers on which the calcification is visible, and - counts the layers counted with the real layer thickness or the layer spacing " x "is multiplied and - the calcifications found in the horizontal sections of the coronary arteries that are longer than ≥ 2 mm are measured once in the layer with the longest dimension and - added to the calcifications calculated in the vertical sections of the coronary arteries.

Description

The invention relates to a method according to the preamble of claim 1.

Ischemic heart disease is a leading cause of death worldwide, according to the World Health Organization (WHO). The early detection of cardiovascular risks is therefore crucial for an extensive prevention of coronary events.

• CAC - Coronary Artery Calcium
• LM - Left Main
• LAD - Left Arterior Descending
• LCX - Left Circum flex Artery
• RCA - Right Coronary Artery

The following abbreviations are used in the following

• CAC - Coronary Artery Calcium
• LM - Left Main
• LAD - Left Arterior Descending
• LCX - Left Circum flex Artery
• RCA - Right Coronary Artery

Changes to the heart and blood vessels, which are usually diagnosed using an elaborate cardiac CT scan, can often be identified on computed tomography of the thorax (thorax CT) initiated for other reasons and can be used to assess or assess the cardiovascular risk profile of a patient. to judge. Typical incidental findings of a chest CT include calcification of the left coronary artery (LAD area), the mitral valve, or the descending aorta. Taken alone, they have no disease value, but they can be signs of atherosclerosis of important blood vessels.

Incidental findings from a chest CT thus indicate cardiac risks.

Current state of research - (see the alphabetically ordered bibliography at the end of the description)

The research group around Shemesh et al. published in 1995 the first study to examine the co-assessment of the heart in low-dose CT examinations of the chest. It was only analyzed whether calcifications were visible on the heart. No comparison was made with standard CT heart examinations, so that the reliability of the method could not be checked. In the years that followed, further studies were added, including those by Wu et al., Budoff et al. and Tacelli et al., who compared low-dose chest exams to cardiac CT exams. In 2008, Wu et al. carried out the first comparative studies. The data from Wu et al. were generated with a 4-line multislice computer tomograph (MSCT). This study was followed by investigations with a 16- and a 64-line device (see Tacelli, N. et al.). Up until then, the CT chest examinations were low-dose examinations from screening and control programs with patients with pulmonary disease or lung cancer patients. However, it showed a very high detectability for CAC.

Patients who are called for pre- or post-operative care after a chest CT therefore benefit from a simple screening method for changes in the coronary arteries, if possible in one examination step without further radiation exposure. All of these studies used low-dose exams, which are the standard for diagnosing lung diseases. Often only a statement about the presence of CAC in the coronary arteries was made in the studies and no CT heart examination was used as a comparison, see e.g. Jacobs, P. et al. 2012

Simple visual scoring systems were typically used for the studies. They were often limited to four categories (0 = none, 1 = mild, 2 = medium and 3 = severe calcification). With these score systems, the coronary arteries were then assessed individually and at the end a total score was created, see Kirsch et al., 2011.

In 2013, Xie et al. published a meta-analysis that compared ten studies with a total of 34,028 patients. In the sub-analysis, five studies with a total of 1316 patients were compared. The pooled correlation between CT chest and CT heart examinations showed a very high agreement of 0.94% between triggered and non-triggered CT examinations. This analysis further examined the mean mortality rate of the studies and came to the conclusion that a positive calcium score in a CT chest examination is associated with an increased mortality rate. The However, the mortality rate is not as high as with an increased CAC score in the CT heart examination, which suggests a certain inaccuracy in the measurement method (Xie et al., 2013).

Comparable studies on this work were carried out by the working groups Azour et al. 2017, Kirsch et al. 2011 and Hutt et al. 2016 created.

Azour et al. published a study in 2017 with a similar rating system as Kirsch et al. 2011. In the CT thorax examinations, the calcifications were classified with a score of zero to twelve. In the CT heart examinations, the AGATSTON CAC score was determined. Azour et al. found a high correlation between the two examination methods (r = 0.811, p <0.001). The negative predictive value for the absence of calcifications was also high (91.6%). The corresponding CAC score values were calculated for the values zero, one to three, four to five and greater than six. A value of zero corresponded to a CAC score of 0.52. A CAC score of 98.7 could be determined for the values one to three. Total values between four and five were assigned a CAC score of 350.6 and from a score of over six the corresponding CAC score was 1925.4. The editors of the study concluded that the visual assessment for both a score of zero and score values above zero from the CT chest examination can predict the corresponding CAC score values well (Azour et al., 2017).

Kirsch et al. compared 2011 cardiac MSCT with thoracic MSCT exams of 163 asymptomatic patients from a whole-body screening program. These two investigations were also carried out in one block without a time delay, as was incidentally also in the context of the present invention. The image acquisition was carried out with a 16- or 64-line CT device. The working group found a high correlation between the two methods of 0.83. A visual score system (Weston score) with a range between zero and twelve was developed to classify the calcifications. Each coronary artery could receive a score between zero and three (zero for no visible calcification, one for minimal calcification, three for calcification that produces blooming and two for calcifications between one and three). The individual results were added up. Here, too, corresponding score values were generated for the clinically relevant CAC score values 100 and 400. These were> 2 and> 7. The working group also found a good correlation with the CT heart examinations for an absence of calcifications in the CT chest examinations. The working group found the best agreement between the two methods for the LAD with a correlation of 0.82 followed by the RCA with 0.62. The LCX (0.55) and the left main stem (0.35) showed poorer matches (Kirsch et al., 2011).

Hutt et al. published a study in 2016 with 185 patients with a positive smoking history. Here, too, the image acquisition took place in an examination block with a 64-line MSCT. The CT heart examination took place prospectively triggered. The same software was used to evaluate the calcium load in the thorax examination as for the evaluation of the CT heart examination. The slice thickness of the thorax examination was 3 mm. The sensitivity for the detection of lime was 96.4%. The specificity was given as 100%. False positive results with regard to the general presence of calcium were not found in any of the investigations, false negative results in only four of the 185 patients. These were in the range of very low CAC scores (0.4 - 2.5). The agreement of the two methods was given as 0.95. The authors of the study found the largest deviations based on the individual evaluation of the arteries for the RCA. They also suggested that this was due to the higher movement of this coronary artery (Hutt et al., 2016).

In the countered US 2019/0 029 625 A1 describes a method with which two-dimensional angiography images can be converted into a 3D data set. Contrast media must be administered for this. Usually these are CT angiography images.

Some studies have also looked at the effects of calcifications seen on chest CT images and the mortality rate. The presence of coronary calculus on chest CT scans has been shown to be a powerful predictor of premature death. Shemesh et al. published the first study on this topic in 2010. They used low-dose chest CT images of 8,782 smokers, who were assigned a score between zero to twelve on the basis of these images. The death rate of the participants was determined using the National Death Index. Cardiovascular deaths increased with an increasing score from the chest CT scan and were 1.2% for a score of zero, 1.8% for a score of one to three, 5.0% for a score of four to six and 5.3% for a score of seven to twelve. If patients with a score of zero were used as the reference group, a score of four was a significant predictor of a cardiovascular event resulting in death (Shemesh et al. 2010).
In 2016, Hughes-Austin et al. published a study comparing 6 mm chest CT with 3 mm cardiac CT scans and relating the results to mortality. They were able to show that, despite the significantly greater layer thickness, there was still a strong correlation between the two imaging techniques and the mortality rate of the study population. In their analysis, they also referred to the high unused resource of this investigation technique (Hughes-Austin et al., 2016).

Vehmas et al. Found a similar relationship with regard to the mortality rate in low-dose CT chest examinations. (2012) and Jacobs et al. (2012).

DESCRIPTION OF THE INVENTION

Determination of the coronary calculus in a CT chest examination

The coronary calculus in the CT thorax examinations was determined with the aid of a computer using software (VivoLab v1.0, Digital Medics, Dortmund, Germany).

The systematic screening of the axial CT thorax images began from the ascending aorta at approximately the level of the exit of the coronary arteries in the standard soft tissue window. The coronary arteries were screened separately for LM, LAD, LCX and RCA.

Each layer in which a point-like calcification ≤ 2 mm could be identified was counted. These punctiform calcifications were particularly noticeable in the vertical parts of the coronary arteries. Particular mention should be made here of segment 2 of the right coronary artery and segments 7, 8 and 10 to 15 of the left coronary artery. Segment 1 of the RCA often contained measurable and countable calcifications. To determine the length of this individual calcification, the number of layers in which the calcification was visible was counted. In order to calculate the total length of the calcification, the counted layers were multiplied by the actual resulting layer thickness (= layer spacing). Since the reconstruction increment (0.75 mm) was chosen to be smaller than the primary slice thickness (1.5 mm), overlapping slices were generated in the CT thorax images used. In the case of the images used, the thickness of an individual slice is therefore 0.75 mm; the slice thickness is set on the CT.

The formula for calculating a single vertically running calcification under 2 mm is therefore: $$\begin{array}{c}\text{Length of a vertical calcification}=\text{Number of layers on which the}\\ \text{Calcification is visible}\times \text{0}\text{, 75 mm layer spacing}\end{array}$$

In this calculation is shown schematically for a layer thickness of 0.75 mm. In calcification is drawn in, which extends over four interlayer spaces. It has a fictional length of 2.25 mm. The calcification lies between layers one to five (Figure 1A). However, it is only trimmed and thus visible in the CT examination in slices two, three and four (Figure 1B). According to the invention, these three layers are evaluated as layers with calcification, counted and multiplied by the layer spacing of 0.75 mm. This results in a calculated length of 2.25 mm.

Figure 1: The calcification is not cut in layers number 1 and 5 (Figure part B).

The calcification extends over layers 2 - 4. Layers 2, 3 and 4 are counted and multiplied by the layer spacing of 0.75 mm. The calculated length of the calcification is 3 x 0.75 mm = 2.25 mm.

If calcification is found in a horizontal section and is longer than ≥ 2 mm, it is measured once in the layer with the longest dimension and added at the end to the further calcifications found ( ). Calcifications ≥ 2 mm can usually be found in segments 3 and 4 of the right coronary artery and in segments 5, 6 and 9 of the left coronary artery, as these tend to run horizontally in the CT images for long stretches (see also and ).

Figure 2: Schematic representation for determining the length of vertical calcifications.

The calcification was measured once in its longest extent.
Here the calcification has a length of 7 mm, for example.

Then all the individual results of an artery are added to the calcifications found in the artery. In order to calculate a total score, the measurements of LM, LAD, LCX and RCA were added at the end (formula 2).
The measurements are rounded to the first place after the decimal point. For each coronary artery, a value with the unit millimeter is created separately. For all coronary arteries together, a total score is obtained after adding all four individual values. The total calcification load of a patient's heart can then be given in millimeters and compared with follow-up examinations or percentiles. $$\begin{array}{l}\mathbf{calculation}prOKOrOnararterie=\text{Length of the calcification / s}\left(\ge 2\text{mm}\right)\\ +\text{Number of layers}\left(\text{at}\le 2\text{mm}\right)\times 0.75\text{mm slice spacing}\\ \mathbf{Partial\; score}prOKOrOnararterie=\text{Sum of the individual calcifications}\\ \mathbf{Total\; score}allerKOrOnararterienderTHOrax-\mathrm{C.}TuntersucHtenPersOn\\ =\text{Sum of the partial scores of the coronary arteries}\end{array}$$

Example images for the evaluation of the coronary calculus in a CT chest examination

The following figures illustrate the methodology for evaluating the CT chest images. The figure part A always shows the CT heart examination, the figure part B the corresponding CT thorax examination on a sectional plane that is as similar as possible.

It should be noted that the coronary arteries differ with regard to the evaluation. This is due to the different course of the coronary arteries, the uneven heart movement and the different spatial expansion of the calcifications (vertical vs. horizontal incision in the CT chest examination). Only the solid calcification is measured without the blooming. During the development of the score, it has been shown empirically that the measurements in both examination methods achieve more similar results if the blooming of calcification is not included (see and ). Blooming is the term used to describe the strong overexposure to CT morphologically particularly dense areas.

With the counting method, however, blooming can be included in the evaluation according to the invention. Since the counting method is used for small calcifications, greater accuracy can be achieved by taking blooming into account. Otherwise, smaller calcifications would not have been included in the evaluation in these areas, as they no longer appear solid.

What is illustrated here using the illustrations as an example corresponds to the majority of the evaluations.

Example images for evaluation on horizontal sections

The LAD runs rather horizontally over long stretches in the axial layers. The calcifications present here can be measured very well and are often congruent to the CT heart examinations. Particularly noteworthy here is segment 6, which is particularly often affected by calcifications (Wasilewski et al. 2015). Movement artifacts also seem to play only a subordinate role here.

and show a comparison of the evaluations in the horizontal portion of the LAD (segment 6). The good agreement of the measurements can be seen. The calcification was measured once in the largest extent according to the score and then added.

Figure 3: Comparison of CT heart (A) and CT thorax examinations (B) of the LAD.

Figure 3A shows calcification measuring 17.8 mm in the LAD. Due to the ECG triggering, this appears more sharply demarcated from the environment than in the CT chest image.
Figure 3B: The 17.8 mm long calcification of the LAD can also be seen, but with clear blurring and blooming (not included in the measurement).

Figure 4: Comparison of CT heart (A) and CT thorax examinations (B) of the LAD.

Figure 4A: Here the calcification in the CT heart examination has a length of 4.9 mm.
Figure 4B: The length of the calcification is 5.87 mm with surrounding blooming, this was not included in the measurement.

Sample images for evaluation on vertical sections

LCX (Left Circumflex Artery) and RCA (Right Coronary Artery) (excluding segments 3 and 4) run more vertically in the axial CT slices over longer sections. Calcifications appear here in dots over several layers and are often smaller than 2 mm, so that they were counted in the method used and not measured. Even in the more vertical segments 7, 8 and 10 of the LAD, the method with counting of the layers was usually indicated ( ; Arrow).

and is intended to illustrate the method using vertical segments. For reasons of clarity, the RCA was chosen as an example. The RCA is particularly susceptible to motion artifact because it moves the fastest during the heart's action. This can lead to problems even with ECG-triggered recordings. If the recordings are not triggered, images result as in Figure 5B and 6B. Here it is advisable to also evaluate artifacts (blooming) as calcifications that arise on the non-ECG-triggered CT thorax examinations. These usually correspond to calcifications in the CT heart examinations and increase the accuracy of the method. In this way, even the smallest lesions can be detected in CT chest images ( (Arrow) and ). The coronary calculus often presents itself as a star-shaped increase in density. According to the score, these calcifications are then simply counted and multiplied by the layer spacing in order to calculate the length.

Figure 5: Comparison of CT heart (A) and CT thorax examinations (B) by the RCA.

Male patient, comparable cutting height, only the heart was shown for reasons of clarity.
Figure 5A shows a clearly delimited calcification in the RCA (ellipse) and, in addition, a small punctiform (<2 mm) calcification in the vertical portion of the LAD (arrow).
The calcification of the RCA can only be seen in Figure 5B as a movement artifact (ellipse); a very small calcification of the LAD can also be seen (arrow). The movement artifact is, however, much less pronounced here than at the RCA. In order to be able to take these calcifications into account in the score, the artifact that appears diffuse in the CT thorax images is also counted without solid components.

Figure 6: Comparison of CT heart (A) and CT thorax examinations (B) by the RCA.

Male patient, comparable cutting height, only the heart was shown for reasons of clarity.
Figure 6A shows a very small but definable calcification in the CT heart examination (arrow); this can only be seen as a very blurred artifact in Figure 6B of the thorax examination (circle). Counted as calcification and therefore counted

uses an inserted stent to illustrate how accurate the counting of layers is for the evaluation. In the stent has a length of 15.2 mm. If the slices in the CT chest imaging on which the stent is visible, including the artifacts, are counted, it can be seen on 21 slices. Multiplied by the layer spacing of 0.75 mm, this corresponds to a dimension of 15.75 mm. So there is only a difference of 0.55 mm. The patient with the stent was chosen for better illustration, as the stent is particularly distinct from the surrounding tissue.

Figure 7: Illustration of the counting method for vertical sections.

The RCA is shown with an inserted stent in MPR (Multiplanar Reformation). The stent has a length of 15.2 mm.

A regression analysis was performed to examine the interrelationships between the two methods, CT heart / CT chest. Because of the right-skewed distribution with many low CAC and CT A logarithmic scale is chosen for thorax score values. Scatter charts are made for graphical illustration.

Percentiles are made for the CAC score and the CT thorax score in order to determine the respective intersections of the relevant CAC score values (CAC score = 100 and CAC score = 400). In order to make a statement about the agreement of the two examination methods, contingency tables are created.

The significance of the investigations was checked using the chi-square test. This is done on the one hand for a CAC score and CT thorax score of zero and a CAC score of 100 and 400 with the respective corresponding CT thorax score.

In order to examine the strength of the influence of the individual risk factors on the results, regression analyzes are preferably carried out. The univariate linear regression examines how each risk factor influences the result. The factors age, BMI, gender, positive smoking history, hypercholesterolemia, hypertension, diabetes mellitus and positive family history were examined. The multivariate regression analysis is used to adjust each individual factor to check the extent to which the previously ascertained relationship can be further confirmed. The same risk factors are taken into account as in the univariate analysis.

For the evaluations of the two analyzes, the odds ratio (OR), the 95% confidence interval and the p-value are preferably included. Box-plot diagrams are preferably made for the risk factors that influenced the evaluation. For all tests, a p-value <0.05 was considered a significant result, a p-value <0.001 a highly significant result.

Results

Quantiles

The CAC and CT thorax scores can be measured in the CT examination for all patients. The upper quantile of the patient collective's CAC score is 168.6. The upper quantile of the CT chest score is 54.6 mm (Table 1). Table 1: Location of the CAC score and the CT chest score with upper quartile, median and 90% quartile. CAC score CT chest score Upper quantile (median; 90%) 168.6 (17.3; 566.1) 54.6 (9.4; 129.8)

Percentiles

The percentile curves of the relevant score values for a CAC score of 100 and a CAC score of 400 intersect at the 68th and 86th percentile, respectively. A CAC score of 100 results in a CT thorax score of 34.3 mm. A CAC score of 400 corresponds to a CT thorax score of 101.7 mm (table). Table 2: Percentile distribution for the CAC score and the CT chest score. 50. 68. 75. 86. 90. 95. CAC score 17.3 100 168.6 400 566.1 1009.5 CT chest score 9.4 mm 34.3 mm 54.6 mm 101.7 mm 129.8 mm 190.7 mm

Correlation of the investigation methods

The analysis of all coronary arteries shows a high correlation of rounded r = 0.94. The strongest correlation in the individual analysis of the coronary arteries is shown in the LAD with rounded r = 0.93. LM (Left Main) and LCX show approximately the same relationship with rounded r = 0.88 (LM) and r = 0.87 (LCX). The correlation for the RCA is 0.83.

Regression analysis

The and describe the results of the regression analysis. The analysis of all coronary arteries shows an explained variance of 98.3%.

• : Regressionsanalyse des gesamten CAC-Scores gegen den gesamten CT-Thorax-Score.

Es ergibt sich die Geradengleichung: $${\text{log}}_{10}\square \left(\text{CAC}-\text{Score}+\text{1}\right)=\mathrm{0,00464}+\mathrm{1,29175}*{\text{<figure-callout id="10" label="log" filenames="DE102019111617B4\_0009.png" state="{{state}}">log</figure-callout>}}_{10}\square \left(\text{CT}-\text{Thorax}-\text{Score}+\text{1}\right).$$

Die erklärte Varianz beträgt 98,3 %.
Mittlere Gerade = Regressionslinie, äußere Geraden = 95 %-Vorhersageintervall.Due to the right-skewed distribution with many low and a few high CAC values, a logarithmic scale was chosen.
The straight line equation results:

• : Regression analysis of the total CAC score against the total CT chest score.

The straight line equation results: $${\text{log}}_{10}\square \left(\text{CAC}-\text{Score}+\text{1}\right)=0.00464+1.29175*{\text{<figure-callout id="10" label="log" filenames="DE102019111617B4\_0009.png" state="{{state}}">log</figure-callout>}}_{10}\square \left(\text{CT}-\text{thorax}-\text{Score}+\text{1}\right).$$

The explained variance is 98.3%.
Middle straight line = regression line, outer straight line = 95% prediction interval.

: Regression analysis of the individual sections of the coronary arteries.
Middle straight line = regression line, outer straight line = 95% prediction interval.

Agreement between the two measurement methods

Table 3 shows the agreement and the deviations of the two examination methods, once in relation to the entire evaluation and then separately for each individual coronary artery. Four groups are formed each time. In two groups the agreement (CAC-Score and CT-Thorax-Score = 0; CAC-Score and CT-Thorax-Score> 0) and in another two groups the deviations (CAC-Score = 0, CT-Thorax-Score) > 0; CAC score> 0, CT thorax score = 0) of the measurements. This also applies to table 4 and table 5 with the corresponding values.For the severity classification, a simplified classification according to RUMBERGER was chosen into only four instead of five groups: zero for no calcifications, greater than zero to 100 for minimal and slight calcifications, greater than 100 to 400 for moderate and greater than 400 for severe calcifications (Rumberger et. Al., 1999).

Overall, 32.7% of patients had a CAC score and a CT chest score of zero. The CAC score and the CT chest score were increased in 63.4% of the patients.

There are deviations between the two examination methods in a total of 4% of the cases with regard to the overall evaluation, whereby it occurs significantly more often (3.7 vs. 0.3%) that the CAC score is positive and the CT thorax score is zero. This pattern is also shown in the individual evaluation of the arteries.

For the left main trunk, 85.9% of the patients tested negative for both the CAC score and the CT chest score. Only 12% of the patients had calcifications in both procedures. The deviation rate was 2.1%. The LAD had negative concurrent scores in 37.0% of patients and positive concurrent scores in 59.6% of patients. There were deviations between the two procedures in only 3.4%. The LCX showed significantly more frequently, namely in 65.9% of the patients a negative CAC score as well as a CT thorax score. In 29.2% of the patients, calcifications were detectable in both procedures. There were deviations in a total of 5% of the cases. The tests on the LCX thus had the highest error rate here. At the RCA, 61.7% of the patients tested negative for both the CT cardiac examination and the CT chest score. 34.5% of the patients had a matching score greater than zero in both methods. There were deviations in both procedures in 3.8% of the evaluations.

The sensitivity was 94.6% and the specificity 99.2%. The positive predictive value was 99.6% and the negative predictive value 90.0%. The p-value was <0.0001. Table 3: Agreement of the measurements with regard to a CAC score and a CT thorax score of zero. CAC score = 0 CT chest score = 0 CAC score = 0 CT chest score> 0 CAC score> 0 CT chest score = 0 CAC score> 0 CT chest score> 0 total 251 (32.7%) 2 (0.3%) 28 (3.7%) 487 (63.4%) LM 660 (85.9%) 3 (0.4%) 13 (1.7%) 92 (12.0%) LAD 284 (37.0%) 4 (0.5%) 22 (2.9%) 458 (59.6%) LCX 506 (65.9%) 2 (0.3%) 36 (4.7%) 224 (29.2%) RCA 474 (61.7%) 7 (0.9%) 22 (2.9%) 265 (34.5%) Table 4 describes the situation in terms of a CAC score of 100 and the corresponding CT chest score of 34.3 mm. Overall, 66.9% of the patients had a CAC score below 100 and a CT chest score below 34.3 mm. In 30.5% of the patients, the CAC score was greater than 100 and a CT chest score greater than 34.3 mm. There were deviations in a total of 2.6% of the cases.

The sensitivity was 96.3% and the specificity 97.9%. The positive predictive value was 95.5% and the negative predictive value was 98.3%. The p-value was <0.0001. Table 4: Agreement of the measurements with regard to a CAC score of 100 and a CT thorax score of 34.3 mm. CAC score <100 CAC score <100 CAC score> 100 CAC score> 100 CT chest score CT chest score CT chest score CT chest score <34.3 mm > 34.3 mm <34.3 mm > 34.3 mm total 514 (66.9%) 9 (1.2%) 11 (1.4%) 234 (30.5%)

Table 5 describes the behavior of the two examination methods for a CAC score of 400 and the corresponding CT thorax score of 101.7 mm. Of the 768 patients examined, 660 (85.9%) had a matching score of <400 in the CAC score and <101.7 mm in the CT chest score. A CAC score of over 400 and a CT chest score of over 101.7 mm were found in 12.4% of the patients. There were deviations in only 1.7% of all examinations.

The sensitivity was 88.0% and the specificity 100%. The positive predictive value was 100.0% and the negative predictive value was 98.1%. With a p-value of <0.0001, this evaluation can also be viewed as statistically significant. Table 5: Agreement of the measurements with regard to a CAC score of 400 and a CT thorax score of 101.7 mm. CAC score < CAC score < CAC score> CAC score> 400 400 400 400 CT chest score CT chest score CT chest score CT chest score <101.7 mm > 101.7 mm <101.7 mm > 101.7 mm total 660 (85.9%) 13 (1.7%) 0 (0%) 95 (12.4%)

Table 6 shows the agreement according to the classification in the respective degrees of severity. The modified classification according to RUMBERGER and the corresponding values for the CT thorax score were selected for this. The data marked in bold show the respective groups with the main agreement. Deviations exist at most in relation to the next higher or next lower group. A total of 8.3% of all patients were classified in the wrong risk group. 6.6% of the patients were classified too low in the CT chest examination, 1.7% too high. Table 6: Correspondences and deviations of the two examination methods with regard to the classification in the corresponding degrees of severity according to RUMBERGER. CT chest score CAC score 0 mm > 0 mm - 34.3 mm > 34.3 mm - 101.7 mm > 101.7 mm total 0 251 32.7% 99.2% 90.0% 2 0.3% 0.8% 0.8% 0 0% 0% 0% 0 0% 0% 0% 253 33.0% > 0 - 100 28 3.7% 10.3% 10.0% 233 30.3% 85.7% 95.5% 11 1.4% 4.0% 7.3% 0 0% 0% 0% 272 35.4% > 100 - 400 0 0% 0% 0% 9 1.2% 6.7% 4.8% 126 16.4% 93.3% 84.0% 0 0% 0% 0% 135 17.6% > 400 0 0% 0% 0% 0 0% 0% 0% 13 1.7% 12.0% 8.7% 95 12.4% 88.0% 100% 108 14.1% total 279 36.4% 244 31.8% 150 19.5% 95 12.4% 768 100%

Influence of risk factors on the evaluation

For the CT thorax score, the influence of risk factors on the evaluability of the examinations was examined. The regression analyzes carried out only showed a significant influence on the measurements for the variables “age” and “BMI” (BodyMassIndex). The patients who were correctly classified in both groups were on average older (59.9 vs. 54.7 years) and had a lower BMI (26.2 vs. 29.0) (Table 7). The further evaluated characteristics, however, did not influence the results (Table 8). Table 7: Factors influencing the evaluation. feature Consistent results Differing results p-value Age 59.9 years 54.7 years 0.0076 BMI 26.2 29.0 0.0011 Table 8: The evaluation non-influencing factors. feature Regarding CAC score and CT chest score p-value consistent results gender • Male 3.9% 1.0 • Female 3.9% Smoking congestion • smokers 4.4% 0.6499 • non-smokers 3.8% Hypercholesterolemia • Positive 4.0% 0.8513 • Negative 3.8% hypertension • Positive 5.1% 0.1863 • Negative 3.1% Diabetes mellitus • Positive 1.9% 0.7155 • Negative 4.1% Family history • Positive 3.7% 1.0 • Negative 4.0%

discussion

Relationship between CAC score and CT chest score

The relationship between the two research methods is confirmed in the statistical analysis. The newly developed CT thorax score shows a high degree of coverage for the CAC score. This is reflected in the high correlation and the high explained variance.

By creating percentiles, it was possible to create corresponding CT thorax values (34.3 mm and 101.7 mm) for the clinically important CAC score values 100 and 400. Using these limit values, it is possible for the clinician to identify patients with an increased cardiovascular risk from chest CT examinations. Due to the very high degree of correspondence between the two examination methods, the CT thorax score also benefits from the good data basis of the CAC score with regard to the classification into the usual clinical risk groups. It is approximately possible to transfer the risk classifications of the CAC score to the CT thorax score. Based on the RUMBERGER severity classification, a CT chest score of zero would correspond to a very low cardiovascular risk. A score greater than zero to 34.3 mm equates to a moderate risk, and a score between 34.3 and 101.7 mm equates to a moderate to high risk. From a value of more than 101.7 mm, the cardiovascular risk is considered to be high (Table 9). Table 9: CAC score and equivalent CT thorax score with a corresponding evaluation of the cardiovascular risk based on this CAC score CT thorax score (in mm) Rating Cardiovascular risk 0 0 No coronary calcifications Very little > 0 - 100 > 0 - 34.3 Mild coronary calcifications Moderate > 100 - 400 34.3 - 101.7 Moderate coronary calcifications Moderate to high > 400 > 101.7 Severe coronary calcifications High

Importance of the absence of calcifications on chest CT images

The absence of calcifications in the CT chest examination has a high negative predictive value for the simultaneous absence of calcifications in the CT heart examination. In 96% of the cases, the two measurement methods agree in the overall evaluation based on a CAC score of zero. The examination according to the invention thus also proves to be suitable for making statements about the health of the heart.

Feasibility of the measurement method

The invention provides a simple method for assessing the calcification load of the heart on a native CT chest examination. No additional software or complex 3D reconstruction is required.

The determination of CAC in the coronary arteries by means of a CT thorax examination is a quick, simple and cost-neutral additional evaluation of a normal CT thorax examination. With a trained eye, the additional evaluation time is two to three minutes with a medium calcium load. By exceeding relevant limit values, a statement can be made about the CAC score that is likely to be expected. The evaluation can be implemented routinely in everyday clinical practice. Furthermore, the development of software for the evaluation is conceivable. The calcification can be measured on any native CT chest examination. The thinner the layer thickness, the more precise results you will probably be able to achieve.

Strengths and weaknesses of the evaluation

The chest CT scan is an examination technique that is not designed to provide an image of the heart. There is also no possibility of eliminating the heart's complex spatial movements. With this examination technique, considerable artifacts are produced in some cases by the movement of the heart. It is not possible to select a specific heart phase because the examination is not triggered by an ECG. Despite all these supposed deficiencies, the method according to the invention can be used to make precise predictions with regard to the CAC score. This is possible if artifacts are correctly assigned, assessed and evaluated.

One limitation is that it is a completely new procedure. There are no comparative values and it is not based on a long-established procedure. Although the method according to the invention has not yet been validated, the high degree of agreement between the two methods has shown that the CAC score can be predicted reliably.

The application of the CT chest score is very easy to perform. It is also not necessary to precisely delimit the individual coronary sections, which would be very difficult anyway with the non-triggered examination technique.

Only the quantity of calcifications is assessed, not the quality.

One of the strengths of the methodology is that there was no time lag between the two examinations. This means that the examination conditions can be assumed to be approximately the same (patient position, progression of calcification, same MTRA). Furthermore, all examinations were carried out with the same CT machine.

It should be noted that all patients had a low heart rate due to the previous medication as a result of the previous CT heart examination. How an increased frequency affects the ability of the images to be assessed is open.

Result

Within the scope of the invention, calcifications of the coronary arteries, which are visible in native CT thorax examinations in the MSCT, were compared with already evaluated CT heart examinations. The peculiarity was that these CT-thorax examinations, in comparison to the CT-heart examinations, were not ECG-triggered. The primary focus was placed on the comparison between this newly developed CT thorax score and the CAC score.

According to the invention, a new measuring method has been developed which does not require additional software and does not have to delimit the individual coronary artery segments. In order to evaluate the CT images effectively and quickly, a method was developed that consists of measuring the calcification and counting the CT slices in which coronary calcification is visible. A score was created with the unit millimeter. The image data came from 768 patients who had received both a CT chest and a CT heart examination in one examination block. The examination was carried out with a 64-line MSCT from Philips.

As a result, a procedure was developed to enable conclusions to be drawn about cardiovascular risks from CT chest examinations.

1. 1. Besteht ein Zusammenhang zwischen den beiden Untersuchungsmethoden?
2. 2. Ab welchem CT-Thorax-Score kommt es zu einem Überschreiten der wichtigen CAC-Score-Werte > 100 und > 400?
3. 3. Welchen negativ prädiktiven Wert hat ein Fehlen von Verkalkungen in CT-Thorax-Bildern?

The answers to three questions were of particular interest when evaluating the test results:

1. 1. Is there a connection between the two examination methods?
2. 2. From which CT thorax score does the important CAC score values> 100 and> 400 occur?
3. 3. What is the negative predictive value of an absence of calcifications in chest CT images?

Re 1 .: The two examination methods showed a high correlation of 0.94 in the overall evaluation. The evaluation of the individual coronary arteries also produced similarly satisfactory results.

Regarding 2 .: By creating percentiles, limit values for the clinically relevant CAC score values of> 100 and> 400 could be formulated. These were 34.3 mm for exceeding a CAC score of 100 and 101.7 mm for exceeding a CAC score of 400.

Regarding 3 .: If no calcifications were visible in the CT chest examination, a CAC score zero was also found in the CT heart examination in about 90% of the cases. A statement about the health of the heart could also be made about the absence of calcifications in the CT chest.

Bibliography on the state of the art (in alphabetical order)

• 1. Azour, L., Kadoch M., Ward T., Eber C., Jacobi, A., (2017): Estimation of Cardiovascular Risk on Routine Chest CT: Ordinal Coronary Artery Calcium Scoring as an Accurate Predictor of Agatston Score Ranges. Journal of Cardiovascular Computed Tomography 11 (1): 8-15. https://doi.org/10.1016/j.jcct.2016.10.001
• 2. Kirsch, J., Buitrago, I., Mohammed, T.-LH, Gao, T., Asher, CR, Novaro, GM, (2011): Detection of coronary calcium during standard chest computed tomography correlates with multi-detector computed tomography coronary artery calcium score. The International Journal of Cardiovascular Imaging, 28 (5), 1249-1256. https://doi.org/10.1007/s10554-011-9928-9
• 3. Hutt, A., Duhamel, A., Deken, V., Faivre, J.-B., Molinari, F., Remy, J., Remy-Jardin, M., (2016): Coronary calcium screening with dual- source CT: Reliability of ungated, high-pitch chest CT in comparison with dedicated calcium-scoring CT. European Radiology, 26 (6), 1521-1528. https://doi.org/10.1007/s00330-015-3978-7
• 4th Agatston, AS, Janowitz, WR, Hildner, FJ, Zusmer, NR, Viamonte, M., Detrano, R., (1990): Quantification of coronary artery calcium using ultrafast computed tomography. Journal of the American College of Cardiology, 15 (4), 827-832. https://doi.org/10.1016/0735-1097(90)90282-T
• 5. Budoff, MJ, Nasir, K., Kinney, GL, Hokanson, JE, Barr, RG, Steiner, R., Casaburi, R., (2011): Coronary Artery and Thoracic Calcium on Non-contrast Thoracic CT Scans: Comparison of Ungated and Gated Examinations in Patients from the COPD Gene Cohort. Journal of cardiovascular computed tomography, 5 (2), 113-118. https://doi.org/10.1016/j.jcct.2010.11.002
• 6th Hughes-Austin, J., Dominguez A., Allison M., Wassel C., Rifkin D., Morgan C., Daniels M., et al., (2016): Relationship of Coronary Calcium on Standard Chest CT Scans With Mortality . JACC. Cardiovascular Imaging 9 (2): 152-59. https://doi.org/10.1016/j.jcmg.2015.06.030
• 7th Jacobs, P., Gondrie, M., van der Graaf, Y., de Koning, H., Isgum, I., van Ginneken, B., Mali, W., (2012): Coronary Artery Calcium Can Predict All- Cause Mortality and Cardiovascular Events on Low-Dose CT Screening for Lung Cancer. American Journal of Roentgenology, 198 (3), 505-511. https://doi.org/10.2214/AJR.10.5577
• 8th. Rumberger, J., Brundage, B., Rader, D., Kondos, G., (1999): Electron Beam Computed Tomographie Coronary Calcium Scanning: A Review and Guidelines for Use in Asymptomatic Persons. Mayo Clinic Proceedings, 74 (3), 243-252. https://doi.org/10.4065/74.3.243
• 9. Shemesh, J., Apter, S., Rozenman, J., Lusky, A., Rath, S., Itzchak, Y., Motro, M., (1995): Calcification of coronary arteries: detection and quantification with doublehelix CT . Radiology, 197 (3), 779-783. https://doi.org/10.1148/radiology.197.3.7480756
• 10. Shemesh, J., Henschke, C., Shaham, D., Yip, R., Farooqi, A., Cham, M., (2010): Ordinal Scoring of Coronary Artery Calcifications on Low-Dose CT Scans of the Chest is Predictive of Death from Cardiovascular Disease 1. Radiology, 257 (2), 541-548. https://doi.org/100.1148/radiol.10100383.
• 11. Tacelli, N., Remy-Jardin, M., Flohr, T., Faivre, J.-B., Delannoy, V., Duhamel, A., Remy, J., (2010): Dual-source chest CT angiography with high temporal resolution and high pitch modes: evaluation of image quality in 140 patients. European Radiology, 20 (5), 1188-1196. https://doi.org/10.1007/s00330-009-1638-5
• 12th Vehmas, T., (2012): Visually scored calcifications in thoracic arteries predict death: follow-up study after lung cancer CT screening. Acta Radiologica (Stockholm, Sweden: 1987), 53 (6), 643-647. https://doi.org/10.1258/ar.2012.120247
• 13th Wasilewski, J., Niedziela, J., Osadnik, T., Duszańska, A., Sraga, W., Desperak, P. Glowacki, J., (2015): Predominant location of coronary artery atherosclerosis in the left anterior descending artery . The impact of septal perforators and the myocardial bridging effect. Kardiochirurgia i Torakochirurgia Polska = Polish Journal of Cardio-Thoracic Surgery, 12 (4), 379-385. https://doi.org/10.5114/kitp.2015.56795
• 14th Wu, M.-T., Yang, P., Huang, Y.-L., Chen, J.-S., Chuo, C.-C., Yeh, C., Chang, R.-S., (2008): Coronary Arterial Calcification on Low-Dose Ungated MDCT for Lung Cancer Screening: Concordance Study with Dedicated Cardiac CT. American Journal of Roentgenology, 190 (4), 923-928. https://doi.org/10.2214/AJR.07.2974
• 15th Xie, X., Zhao, Y., Bock, GH de, Jong, PA de, Mali, WP, Oudkerk, M., Vliegenthart, R., (2013): Validation and Prognosis of Coronary Artery Calcium Scoring in Nontriggered Thoracic Computed Tomography Clinical Perspective. Circulation: Cardiovascular Imaging, 6 (4), 514-521. https://doi.org/10.1161/CIRCIMAGING.113.000092

Claims (6)

Method for determining / evaluating / assessing the cardiovascular risk profile of a person by evaluating a thorax computed tomography of this person, characterized in that to determine the total length of a calcification - each thorax CT slice of the vertically extending portions of the LM, LAD, LCX - and RCA coronary arteries, in which a point-like calcification ≤ 2mm is identified, counts, and - counts the number of layers on which the calcification is visible, and - the counted layers with the real layer thickness or the layer spacing "x" multiplied and - the calcifications found in the horizontal sections of the coronary arteries that are longer than ≥ 2 mm are measured once in the layer with the longest size and - added to the calcifications calculated in the vertical sections of the coronary arteries. Procedure according to Claim 1 , characterized in that the layer spacing is x ≤ 1 mm. Procedure according to Claim 2 , characterized in that the layer spacing is x ≤ 0.75 mm. Procedure according to Claim 1 , characterized in that in the vertically extending portions of the coronary arteries, increased attention is paid to segment 2 of the right coronary artery and segments 7, 8 and 10 to 15 of the left coronary artery. Procedure according to Claim 1 , characterized in that, in the horizontal sections of the coronary arteries, increased attention is paid to calcifications ≥ 2 mm in segments 3 and 4 of the right coronary artery and in segments 5, 6 and 9 of the left coronary artery. Method according to one of the Claims 1 to 5 , characterized in that the total calcification load of the heart is calculated in mm according to the following formula and compared with subsequent cardiac CT examinations or percentiles of the person in question: - Calculation per coronary artery = length of calcification / s (≥ 2mm) + number of the slices (at ≤ 2mm) x mm slice distance - partial score per coronary artery = sum of the individual calcifications - total score of all coronary arteries of the thorax CT of the person examined = sum of the partial scores of the coronary arteries

Images