Best of the Best Oral Presentation Clinical Oncology Society of Australia Annual Scientific Meeting 2020

The International Lung Screening Trial: opportunistic screening for osteoporosis using low-dose chest CT; baseline findings (#31)

Nikita Botadra 1 , Katrina Hopcraft 2 3 , Rachael O'Rourke 2 3 , Anne Williamson 2 3 , Paraskevi Georgiou 2 , Kwun Fong 1 2 , Henry Marshall 1 2
  1. Thoracic Medicine Department, The Prince Charles Hospital, Brisbane, Queensland, Australia
  2. Thoracic Research Centre, University of Queensland, Brisbane, Queensland, Australia
  3. Medical Imaging Department, The Prince Charles Hospital, Brisbane, Queensland, Australia

Aim: Screening with low-dose CT (LDCT) can reduce lung cancer mortality(1). Detection of comorbidities which share risk factors with lung cancer, such as osteoporosis (OP), may add value.  Vertebral bone mineral density (BMD) and vertebral fractures (VF) can be measured on LDCT. We describe the baseline prevalence of OP at a single site of the ILST lung cancer screening trial using two BMD analysis methods in comparison to validated clinical risk tools (FRAX and Garvan calculators).

Method: 595 participants aged 55-80 at high risk of lung cancer underwent baseline LDCT at The Prince Charles Hospital ILST site (Queensland)(2). A single trained reader analysed scans for 1) VF (semiquantitative Genant method), 2) BMD using two methods [L1 vertebra Hounsfield Unit attenuation value (L1HU) and gold-standard quantitative CT (Mindways software, ‘QCT’)].  Low BMD thresholds were defined as <110HU and <120mg/cm3 respectively. OP fracture risk scores were calculated from questionnaires.

Results: 492 complete cases were analysed (mean age 64.9, 59% male). 131/492(27%) had a high 10-year fracture risk (FRAX or Garvan, probability >3% of hip or >20% of other major OP fracture). 170/492(35%) had moderate-severe VF (>25% loss of height) .

290/492(59%) had low lumbar BMD defined by QCT, compared to 205/492(42%) defined by L1HU. Compared to QCT, classification of low BMD by L1HU was excellent (AUROC 0.945(95%CI:0.93-0.96). Both methods classified prevalent VF equally well (AUROC L1HU and QCT 0.636 and 0.654 (p= 0.17)). Both clinical risk scores had statistically significantly lower performance than BMD (AUROC FRAX 0.495(p= 0.00167), Garvan 0.493(p= 0.001462)).

Conclusion: In this lung cancer screening cohort, VF and low BMD were highly prevalent. Both measures of BMD out-performed risk scores in predicting OP (i.e. prevalent moderate-severe VF). The L1HU method, validated against gold-standard, can reliably and quickly be applied to screening LDCT to opportunistically detect occult OP in high-risk smokers.

  1. National Lung Screening Trial Research Team. Reduced lung-cancer mortality with low-dose computed tomographic screening. New England Journal of Medicine. 2011 Aug 4;365(5):395-409.
  2. Lim KP, Marshall H, Tammemägi M, Brims F, McWilliams A, Stone E, Manser R, Canfell K, Weber M, Connelly L, Bowman RV. Protocol and Rationale for the International Lung Screening Trial. Annals of the American Thoracic Society. 2020 Apr;17(4):503-12.