QCT, the most accurate method of measuring bone mineral density?

Opportunistic AI-enabled automated bone mineral density measurements in lung cancer screening and coronary calcium scoring CT scans are equivalent

Rationale and objectives

We previously reported a novel manual method for measuring bone mineral density (BMD) in coronary artery calcium (CAC) scans and validated our method against Dual X-Ray Absorptiometry (DEXA). Furthermore, we have developed and validated an artificial intelligence (AI) based automated BMD (AutoBMD) measurement as an opportunistic add-on to CAC scans that recently received FDA approval. In this report, we present evidence of equivalency between AutoBMD measurements in cardiac vs lung CT scans.

Materials and methods

AI models were trained using 132 cases with 7649 (3 mm) slices for CAC, and 37 cases with 21918 (0.5 mm) slices for lung scans. To validate AutoBMD against manual measurements, we used 6776 cases of BMD measured manually on CAC scans in the Multi-Ethnic Study of Atherosclerosis (MESA). We then used 165 additional cases from Harbor UCLA Lundquist Institute to compare AutoBMD in patients who underwent both cardiac and lung scans on the same day.

Results

Mean±SD for age was 69 ± 9.4 years with 52.4% male. AutoBMD in lung and cardiac scans, and manual BMD in cardiac scans were 153.7 ± 43.9, 155.1 ± 44.4, and 163.6 ± 45.3 g/cm³, respectively (p = 0.09). Bland-Altman agreement analysis between AutoBMD lung and cardiac scans resulted in 1.37 g/cm³ mean differences. Pearson correlation coefficient between lung and cardiac AutoBMD was R² = 0.95 (p < 0.0001).

Conclusion

Opportunistic BMD measurement using AutoBMD in CAC and lung cancer screening scans is promising and yields similar results. No extra radiation plus the high prevalence of asymptomatic osteoporosis makes AutoBMD an ideal screening tool for osteopenia and osteoporosis in CT scans done for other reasons.

Liver Fat Content Measurement with Quantitative CT Validated against MRI Proton Density Fat Fraction: A Prospective Study of 400 Healthy Volunteers

Background Although chemical shift-encoded (CSE) MRI proton density fat fraction (PDFF) is the current noninvasive reference standard for liver fat quantification, the liver is more frequently imaged with CT. Purpose To validate quantitative CT measurements of liver fat against the MRI PDFF reference standard. Materials and Methods In this prospective study, 400 healthy participants were recruited between August 2015 and July 2016. Each participant underwent same-day abdominal unenhanced quantitative CT with a calibration phantom and CSE 3.0-T MRI. CSE MRI liver fat measurements were used to calibrate an equation to adjust CT fat measurements and put them on the PDFF measurement scale. CT and PDFF liver fat measurements were plotted as histograms, medians, and interquartile ranges compared; scatterplots and Bland-Altman plots obtained; and Pearson correlation coefficients calculated. Receiver operating characteristic curves including areas under the curve were evaluated for mild (PDFF, 5%) and moderate (PDFF, 14%) steatosis thresholds for both raw and adjusted CT measurements. Sensitivity, specificity, positive predictive value, and negative predictive value were calculated. Results Four hundred volunteers (mean age, 52.6 years ± 15.2; 227 women) were evaluated. MRI PDFF measurements of liver fat ranged between 0% and 28%, with 41.5% (166 of 400) of participants with PDFF greater than 5%. Both raw and adjusted quantitative CT values correlated well with MRI PDFF (r2 = 0.79; P < .001). Bland-Altman analysis of adjusted CT values showed no slope or bias. Both raw and adjusted CT had areas under the receiver operating characteristic curve of 0.87 and 0.99, respectively, to identify participants with mild (PDFF, >5%) and moderate (PDFF, >14%) steatosis, respectively. The sensitivity, specificity, positive predictive value, and negative predictive value for unadjusted CT was 75.9% (126 of 166), 85.0% (199 of 234), 78.3% (126 of 161), and 83.3% (199 of 239), respectively, for PDFF greater than 5%; and 84.8% (28 of 33), 98.4% (361 of 367), 82.4% (28 of 34), and 98.6% (361 of 366), respectively, for PDFF greater than 14%. Results for adjusted CT were mostly identical. Conclusion Quantitative CT liver fat exhibited good correlation and accuracy with proton density fat fraction measured with chemical shift-encoded MRI. © RSNA, 2019 Online supplemental material is available for this article.

Simultaneous screening for osteoporosis at CT colonography: bone mineral density assessment using MDCT attenuation techniques compared with the DXA reference standard

The purpose of this study was to evaluate the utility of lumbar spine attenuation measurement for bone mineral density (BMD) assessment at screening computed tomographic colonography (CTC) using central dual-energy X-ray absorptiometry (DXA) as the reference standard. Two-hundred and fifty-two adults (240 women and 12 men; mean age 58.9 years) underwent CTC screening and central DXA BMD measurement within 2 months (mean interval 25.0 days). The lowest DXA T-score between the spine and hip served as the reference standard, with low BMD defined per World Health Organization as osteoporosis (DXA T-score ≤ -2.5) or osteopenia (DXA T-score between -1.0 and -2.4). Both phantomless quantitative computed tomography (QCT) and simple nonangled region-of-interest (ROI) multi-detector CT (MDCT) attenuation measurements were applied to the T(12) -L(5) levels. The ability to predict osteoporosis and low BMD (osteoporosis or osteopenia) by DXA was assessed. A BMD cut-off of 90 mg/mL at phantomless QCT yielded 100% sensitivity for osteoporosis (29 of 29) and a specificity of 63.8% (143 of 224); 87.2% (96 of 110) below this threshold had low BMD and 49.6% (69 of 139) above this threshold had normal BMD at DXA. At L(1) , a trabecular ROI attenuation cut-off of 160 HU was 100% sensitive for osteoporosis (29 of 29), with a specificity of 46.4% (104 of 224); 83.9% (125 of 149) below this threshold had low BMD and 57.5% (59/103) above had normal BMD at DXA. ROI performance was similar at all individual T(12) -L(5) levels. At ROC analysis, AUC for osteoporosis was 0.888 for phantomless QCT [95% confidence interval (CI) 0.780-0.946] and ranged from 0.825 to 0.853 using trabecular ROIs at single lumbar levels (0.864; 95% CI 0.752-0.930 at multivariate analysis). Supine-prone reproducibility was better with the simple ROI method compared with QCT. It is concluded that both phantomless QCT and simple ROI attenuation measurements of the lumbar spine are effective for BMD screening at CTC with high sensitivity for osteoporosis, as defined by the DXA T-score.

Osteoporosis in a female population from Bratislava--age-related BMD changes

PATIENTS AND METHODS

We analysed 498 women (n=498) in a Bratislava (BA) population aged 21 to 90. We measured bone mineral density (BMD) in the proximal femur with one densitometric instrument (DXA Osteocore II, France; dual energy X-ray absorptiometry), applying BMD and T-score values in three standard regions of interest: Neck (ROI1), Ward's area (ROI2), Trochanter (ROI3).

RESULTS

Measured values of T-score in ROI1, ROI2 had normal distribution and a lognormal distribution of frequency in ROI3. Using chi2-test (chi-square goodness-of-fit statistics), we determined the distribution of the frequency of T-score values and the percentage of osteoporosis incidence in the Bratislava female population. The osteoporosis incidence, according to T-score values measured in ROI1 was 2.40%, in ROI2 16.34% and in ROI3 3.83%. Following the division of women into ten-year intervals, the statistically significant sample averages of T-score values were decreasing in relation to age only for ROI2. Osteoporosis incidence in age intervals was rising with age for ROI2, for ROI1 the number of osteoporotic patients in the 61 to 70-year interval was lower than in the 41 to 50-year interval, and for ROI3 the number of osteoporotic patients in the 51 to 60-year interval was lower than in the 41 to 50-year interval. Except in the above-mentioned intervals, T-score values decreased in relation to age also in ROI1 and ROI3. According to the analysis of variance, the age category explains 9.6% of the overall variability of T-score values for ROI1, 24.7% for ROI2 and 11.70% for ROI3.

CONCLUSIONS

As in ROI2 (Ward's area) a greater fraction of trabecular bone is measured in comparison with ROI1 and ROI3, ROI2 reflects best the age-related BMD changes. In ROI1 and ROI3 the relation was distorted by a greater fraction of cortical bone in comparison with ROI2 and by osteoarthritis.

Phalangeal quantitative ultrasound measurements in former pre-term children aged 9-11 years

The objective of this study was to compare phalangeal ultrasound values in 38 former pre-term children, aged 9-11 years, with 50 age-matched term controls. Skeletal status was evaluated using phalangeal quantitative ultrasound measurements (QUS) by DBM Sonic 1200 (IGEA, Carpi, Italy) which measures the amplitude dependent speed of sound (Ad-SoS, m s(-1)). There were no significant differences in values of Ad-SoS, weight and height between patients and controls irrespective of birth weight or prematurity. In conclusion, phalangeal ultrasound measurements performed in prematurely born infants show that at the age of 9-11 years their bone status does not differ from children born at term.

Screening for lung cancer with low-dose spiral computed tomography

Studies suggest that screening with spiral computed tomography can detect lung cancers at a smaller size and earlier stage than chest radiography can. To evaluate low-radiation-dose spiral computed tomography and sputum cytology in screening for lung cancer, we enrolled 1,520 individuals aged 50 yr or older who had smoked 20 pack-years or more in a prospective cohort study. One year after baseline scanning, 2,244 uncalcified lung nodules were identified in 1,000 participants (66%). Twenty-five cases of lung cancer were diagnosed (22 prevalence, 3 incidence). Computed tomography alone detected 23 cases; sputum cytology alone detected 2 cases. Cell types were: squamous cell, 6; adenocarcinoma or bronchioalveolar, 15; large cell, 1; small cell, 3. Twenty-two patients underwent curative surgical resection. Seven benign nodules were resected. The mean size of the non-small cell cancers detected by computed tomography was 17 mm (median, 13 mm). The postsurgical stage was IA, 13; IB, 1; IIA, 5; IIB, 1; IIIA, 2; limited, 3. Twelve (57%) of the 21 non-small cell cancers detected by computed tomography were stage IA at diagnosis. Computed tomography can detect early-stage lung cancers. The rate of benign nodule detection is high.

The use of clinical CT for baseline bone density assessment

PURPOSE

Many patients having an abnormal initial bone densitometry study have had a previous abdominal/pelvic computed tomography (CT) for other clinical reasons. This study evaluates if a nondedicated quantitative CT (QCT) abdominal/pelvic CT scan could be used as reliable baseline data for subsequent dedicated bone density studies.

SUBJECTS AND METHODS

Twenty-six patients (13 men, 13 women) undergoing clinically-indicated non-i.v. and i.v. contrast abdominal/pelvic CT had dedicated QCT performed immediately following scans of the L1, L2, and L3 vertebral bodies. QCT was then performed on all three scans. A repeated measures analysis of variance model was used to analyze the data in order to compare noncontrast clinical CT with QCT and noncontrast clinical with contrast clinical CT.

RESULTS

The mean bone mineral density for the noncontrast clinical study was 98.51 (mg/cc) versus 90.56 (mg/cc) for QCT (p = 0.0003; 95% confidence interval: 3.90 to 13.71). There was no significant difference (p = 0.085) between QCT performed from non-i.v. and i.v. contrast clinical CT scans.

CONCLUSION

Bone densitometry can be performed from either non-i.v. or i.v. contrast clinical CT scans if a conversion factor is applied. This can be determined by utilizing a formula Daverage = -7.83 + (0.99 x NCaverage), where Daverage and NCaverage are the abbreviations of "dedicated" and "noncontrast clinical" BMD averaged over vertebral bodies L1-L3, respectively.