What’s new in Cardiothoracic Imaging – January 2021

1 year ago

Analysis of the 30-Pack-Year Smoking Threshold in African Americans From an Underserved Lung Cancer Screening Program

Basu, A., Kopulos, L., Geissen, N., Sukhal, S., & Smith, S. B. (2020).

Journal of the American College of Radiology. doi:10.1016/j.jacr.2020.08.017

In an article written from Cook County and Northwestern University Hospitals, researchers described downfalls in the 30-pack-year-smoking (PYS) model for lung cancer screening (LCS) in the African American (AA) community. Between 2017 and 2019, 827 patients were referred for LCS with 784 patients (948%) meeting smoking history criteria. LCS was performed by USPTF and ACR best practices using Low Dose Computed Tomography (LDCT) read by board certified radiologists who utilize the Lung CT Screening Reporting and Data System (Lung-RADS). Lung-RADS 1 and 2 results were considered negative exams and 3 and 4 were considered positive, were referred to a multidisciplinary lung nodule clinic, biopsy and ultimately treatment if needed. Still-smokers were referred for smoking cessation.

Enrolled patients self-reported their PYS and packs smoked per day (PPD). Patients were predominately male (77.5%) and African American (66.2%). The median age of the study population was 62 years. The median years smoked was 40 with an inner quartile range (IQR) of 30-45 years. The median PPD= 1 (IQR 0.5-1) and median PYS was 25 years (IQR of 15-40). 75.5% of patients were still-smokers while 55% of patients reported smoking less than 30 pack years. It was noted that AA patients had the same number of years-smoked (40 [IQR 30-45] P= .252), but had fewer PYS (24 [IQR 15-40] P<.001) compared to other groups due to a lower PPD reported (0.75 [IQR .5-1.0] P<.001).

In round one of LCS the majority of patients had either Lung-RAD 1 (57.6%) or 2 (36.0%) scores while 5.7% received a 3 or 4, respectively. The lung cancer incidence overall in the trial was 2.1% for PYS>30 and 2.0% for PYS<30. Subgroup analysis revealed the lung cancer incidence for AA was 2.5%, although this was not significantly increased from their white counterparts (1.9%, P value= 0.598); it should be noted that the AA patients diagnosed with lung cancer all had less than 30 pack years smoked. It should also be noted that there was no significant differences in cancer incidence in AA patients between those with <30 PYS and >30 PYS. During regression analysis it was observed that female gender (OR 10.57 95% CI 2.4-46.3, P=.007) and co-diagnosis of COPD (OR 4.26 95% CI 1.2-14.99, P=.025) correlated with increased incidence of lung cancer.

This study outlined nicely why LCS guidelines may need further inspection and updating. The NELSON trial for which the current USPTF guidelines are based only had a 4% AA patient population enrolled in their trial. This study showed that AA patients report fewer pack-years smoked than other populations and that there is no differences in incidence between those who have >30 PYS and those who have <30 PYS. Review of screening guidelines to include AA patients who have less than 30 PYS may help in identifying a population who are disproportionately harmed by lung cancer. Of note, there is limited data on long term mortality given the relatively short study interval. This may be an interesting place for further research to focus. This was also a relatively small sample size and as such small differences in study populations are not found to be statistically significant. Repeating this study on a larger scale may provide further detail.

 

Detection of Extrapulmonary Malignancy During Lung Cancer Screening: 5-Year Analysis at a Tertiary Hospital

Chintanapakdee, W., Mendoza, D. P., Zhang, E. W., Botwin, A., Gilman, M. D., Gainor, J. F., . . . Digumarthy, S. R. (2020).

Journal of the American College of Radiology, 17(12), 1609-1620. doi:10.1016/j.jacr.2020.09.032

Researchers from Massachusetts General Hospital analyzed the number of malignancies diagnosed on incidental lesions seen on Low-Dose CT (LDCT) scans for lung cancer screening (LCS) and their associated costs. This study involved retrospective review of 7414 scans for 4160 patients to identify 3165 scans with the “S” modifier per the Lung CT Screening Reporting and Data System (Lung-RADS). 303 of these scans were then whittled down to a final cohort of 229 scans in 225 patients with 241 total lesions; those excluded were done so on the basis of being previously reported, known malignancy, or those who were lost to follow up. These 241 lesions were then worked up via standard protocols and malignancy or benignity was determined by final tissue pathological diagnosis. The lesions were stratified by location into twelve subcategories including: Thyroid, Kidney, Adrenal, Breast, Mediastinum, Intrathoracic lymph node, Chest wall (including axilla), liver, other upper abdominal, bone, esophagus and pleura. Associated costs with the indeterminate lesions’ workup were estimated using the 2020 Current Procedural Terminology Codes and total facility relative value units (RVU’s) for each organ type. Out of pocket expenses were tracked by patient billing records.

The mean study age was 66.1 years and the study was equally balanced by gender (109 male:116 female). Incidental lesions were most commonly found within the Thyroid (23.7%), Kidney (20.8%), Adrenal gland (10.8%) and Breast (10.0%). Of the 241 lesions, 193 underwent further workup. Final diagnosis of benignity was established in 207/241 patients after further workup. 20.3% of indeterminate lesions underwent tissue sampling, which resulted in a malignant diagnosis in 20 lesions. True positives and False positives were calculated for each organ type, allowing for the estimation of an organ-specific positive predictive value (PPV) and negative predictive value (NPV). Costs associated with workup were obtained and estimates were calculated per lesion and per patient.

Extra pulmonary malignancy was detected in 20/241 lesions for a prevalence of 8.9% in those with indeterminate lesions and 0.48% in the general population (20/4,160 patients imaged). The PPV was highest for lesions within the chest wall and axilla (36.4%), Bone (25%) and Breast (12.5%). The total cost of working up all indeterminate lesions was $26,320.52, while the average cost per additional malignancy was calculated at $1,360.03 ($26,320/20 malignancies), $6.33 per participant ($26,320.52/4,160 participants) and $109.21 of excess cost per indeterminate lesion ($26,320.52/241 lesions). 90% of patients (203/225) had no out-of-pocket costs, while those that did had an average cost of $160.60. In subsequent analysis it was found that lesions in organs with low PPV’s, such as the Thyroid had increased costs associated with the diagnosis of malignancy when compared to organs that higher PPV’s such as the bone sub group ($2,783.50:$264.18).

The prevalence of extra pulmonary malignancy found incidentally in the lung cancer screening population mirror that of the COSMOS trial, with a relative increase observed in the rate of incidence in the population with indeterminate lesions (8.9% compared to 6.2%). This was explained by studies being read entirely by thoracic-trained radiologist and improved iterative reconstruction techniques that are now available. The cost per participant ($6.33) and cost per indeterminate lesion ($109.21) are quite low for the added benefit of catching early, treatable malignancy. While the low negative predictive value observed in certain organ systems and complete lack of malignancy identified in others raises important questions over what should be done when incidental lesions are detected in those areas. This is a suggested area of further research.

 

Dual-Energy CT Pulmonary Angiography (DECTPA) Quantifies Vasculopathy in Severe COVID-19 Pneumonia

Radiology: Cardiothoracic imaging; vol. 2 no 5. Oct 29, 2020.

Ridge, C, Desai, S, et al.

Researchers from the Royal Brompton Hospital in London, UK sought to investigate the utility of dual energy CT pulmonary angiography (DECTPA) to evaluate the influence of COVID-19 and the virus’s role on disease duration, RV dysfunction, lung compliance, D-dimer, and obstruction index. The quantitative variable investigated on DECTPA was relative perfused blood volume (pulmonary blood volume/pulmonary artery enhancement, or PBV/PAenh). An automated generated color scale was created using enhancement patterns; perfusion defects were determined qualitatively by observing segmental hypoenhancement using this color scale. Study participants included 27 ventilated patients who underwent DECTPA to diagnose pulmonary thrombus, with 11 surveillance patients who scanned roughly 2 weeks later.  Patients were excluded if BMI >35, as well as patients unable to position arms above head in order to reduce beam hardening artifacts.  Images were acquired with a trigger of 120 HU in the main pulmonary artery with PA enhancement being recorded with HU/sec using bolus tracking CT images. Each voxel was derived from a 3-material decomposition algorithm for air, soft tissue, and iodine and was reconstructed using a dense lung PBV map with a HU threshold of -200. Morphology of perfusion defects (wedge shaped, mottled, or amorphous), presence of thrombus within the PA, and the order of involvement along the pulmonary artery tree were evaluated to see if there was anatomical correlation. The control participants used for this study were obtained from examinations used from a prior case control study at the same facility.

Filling defects consistent with PE were seen in 11 of 27 patients, with similar distribution through each order of the pulmonary arterial tree.  Visually scored perfusion defects were seen in all DECTPA studies, with the amorphous pattern of defect the most prominent (n=21). The predominant pattern of perfusion defect was only attributed to PE in 2/27 cases.  PBV/PAenh in COVID-19 patients was shown to be significantly decreased when compared to healthy controls (17.5%±4.4% vs 27±13.9% (p=0.002)).  RV dysfunction diagnosed by echocardiography was present in 33/38 cases, however was not shown to be related to differences in PA enhancement.  There was an inverse relationship of PBV/PAenh and RV dysfunction when controlling for age, gender, BMI, and ethnicity (B= -5.6, SE 1.6, p <0.001). When comparing symptom onset to timing of initial DECTPA, there was no statistical correlation to average perfusion defect and duration of disease, but when the study was obtained <14 days from onset, visually scored perfusion defects were greater than after 14 days (40.6% vs 26.3% after 14 days, p <0.05).  Overall, the results show that in patients with COVID-19 respiratory failure,  there is reduced pulmonary relative perfused blood volume as expected, and that over time PBV/PAenh improves suggesting that pulmonary angiopathy may reflect an acute feature of severe COVID-19 pneumonia.  This imaging finding is potentially useful as an important marker and surveillance tool during the duration of the COVID-19 pandemic.

There are multiple limitations with this study. There was a small cohort size, the participants all came from a single center, a single scanner was used for all acquisitions, and pathology assumptions are made from imaging without histologic proof (although biopsy for histological proof seems excessive in this patient population).  Additional studies are needed comparing the findings of respiratory failure with DECTPA in COVID-19 to other pulmonary infections, as well as other etiologies of respiratory failure.

 

Evaluation of a Tube Voltage– Tailored Contrast Medium Injection Protocol for Coronary CT Angiography: Results From the Prospective VOLCANIC Study

Schoepf, U, Martin, S, et. al.

American Journal of Roentology Issue 215, Nov 2020

Researchers from the Medical University of South Carolina sought to investigate the feasibility of commercially available software to determine if tube voltage tailored contrast medium dosage would be useful in reducing contrast load and radiation dose.  120 patients were selected for the study and were evenly divided amongst 7 groups with increasing kV values ranging from 70-130 kV and corresponding contrast doses ranging from 33 mL to 65 mL, respectively.  Scans were obtained prospectively with ECG-triggered adaptive sequential mode acquisition using Siemens CT scanners.  Readers with 2-4 years of cardiovascular imaging experience rated the studies subjectively on a scale of 1-5 (1 worst, 5 best) in regards to contrast enhancement, image noise, and overall image quality.  HU attenuation was measured in the aorta, main pulmonary artery, left main (LM) and right coronary arteries (RCA)  as well as left anterior descending (LAD) and left circumflex coronary (LCx) arteries,

Results show that contrast attenuation in the arteries peaked in the 70 kV group, while it was lowest in the 110 and 130 kV groups.  No significant differences were observed in signal-to-noise ratio and in contrast-to-noise ratio between the kV cohorts.  The 70 kV group was shown to have a contrast-to-noise ratio figure-of-merit (CNR-FOM) significantly different from the remaining tube voltages.  CNR-FOM is a calculated value obtained from dividing the square of the CNR by radiation effective dose.  Overall image quality was rated as diagnostic across all tube voltages by the specialized readers. Overall, the results show that dedicated software is able to determine patient specific contrast media administration to reduce overall total contrast dose and radiation dose.  Specifically, lower doses such as at 70 kV can be used for study acquisition without decreasing subjective or objective image quality.

There are limitations to this study.  This study only investigated image quality and did not attempt to evaluate the effect of the tube voltage difference on pathology, such as coronary artery stenosis.  Software used to select the tube voltage is only available from a single vendor and thus broad market acceptance would be difficult.  Finally, the acquisition groups did not all have the same amount of patients included which could affect the representation of certain tube voltages.

 

In Situ Pulmonary Artery Thrombosis: Unrecognized Complication of Radiation Therapy

Wu, C, Ahuja, J, et. al.

American Journal of Roentology Issue 215, Nov 2020

Researchers from MD Anderson Cancer Center  evaluated the role of radiation therapy with in situ pulmonary artery thrombosis.  Cardiovascular complications are the leading cause of non-cancer related morbidity and mortality in patients who undergo radiation therapy, which include accelerated coronary artery atherosclerosis, valvular sclerosis, myocardial disease, pericardial disease, conduction disease, and autonomic diseases. In situ pulmonary artery thrombosis is when the thrombus develops in the pulmonary artery, and not a migration from a deep vein source.

27 patients were determined to have radiation therapy induced in -situ thrombosis by retrospective analysis of the imaging database .  Patients were excluded if the thrombus was deemed tumor or stump thrombus following a lobectomy and/or pneumonectomy.  Multiple variables were examined including: central vs eccentric thrombus, PA segment involved, occlusive vs non-occlusive, angle of thrombus to the vessel wall, presence of lung fibrosis, right heart strain, and relationship of thrombus with radiation therapy (RT).  Review of imaging showed that RT induced thrombus was more common to be a single nonocclusive eccentric thrombus with an obtuse angle in relation to the vessel wall.  The thrombi were exclusively located within the pulmonary vasculature affected by radiation changes.  There was no single artery location (main, lobar, or segmental) that showed a predilection to formation, and the affected vessel was based solely on where the radiation induced fibrosis had occurred.  This study is important in developing the literature regarding RT induced in-situ PA thrombosis, as the clinical and treatment implications of RT induced thrombus are unknown.  Review of CT examinations for this study did not show embolization to the ipsilateral or contralateral lung on follow up imaging.  This could play an important role in determining if anticoagulation is needed with this condition.  One important consideration is for radiologists knowing this entity and being able to distinguish it from acute PE.

This study is a good initial start, however every study has limitations.  This study does have a small study size and a larger patient study population could have different results. Additionally, there is a lack of pathologic confirmation of the imaging findings.  Further larger powered studies are needed to determine the differences between acute pulmonary thromboembolism and radiation induced in situ pulmonary artery thrombosis.

References
  • Share