Jump to comment:

• Is the use of the Herder score valid in these guidelines?
  Deborah R Pencharz, Anant Patel and Thomas L Wagner
  Published on: 28 October 2020

• Published on: 28 October 2020

  Is the use of the Herder score valid in these guidelines?

  • Deborah R Pencharz, Consultant in Nuclear Medicine Royal Free London NHS Foundation Trust
  • Other Contributors:
    • Anant Patel, Consultant in Respiratory Medicine
    • Thomas L Wagner, Consultant in Nuclear Medicine

  Dear Editors,
  This journal published the BTS guidelines for the management of pulmonary nodules in August 2015 (1), leading to widespread evidence-based management of this common clinical problem. The use of the Herder score (2) to estimate or predict the likelihood of malignancy has since become routine in lung cancer MDMs around the country.
  We therefore wish to highlight that the Herder prediction model was developed using the intensity of FDG uptake (absent, faint, moderate or intense) from the uncorrected PET images. However, as far as we are aware, lung cancer MDMs routinely assess the intensity of FDG uptake from the corrected images which is not in accordance with the original Herder model.
  The use of uncorrected images in the original Herder study (2) to distinguish between uptake categories potentially alters the perceived evidence base of the recommendations in the BTS guidelines (1) to distinguish between faint and moderate uptake according to mediastinal blood pool given that this scale of uptake was not used in the original score, has not been validated and could mean we are not using the correct category in the risk model.
  Furthermore, when considering the widespread use of the Herder score, it should be appreciated that it was formulated from patients scanned between 1997 and 2001. The Herder paper describes that “emission scans were acquired in a two-dimensional mode … and were reconstructed using ordered subset expectation maximisa...

  Show More

  Dear Editors,
  This journal published the BTS guidelines for the management of pulmonary nodules in August 2015 (1), leading to widespread evidence-based management of this common clinical problem. The use of the Herder score (2) to estimate or predict the likelihood of malignancy has since become routine in lung cancer MDMs around the country.
  We therefore wish to highlight that the Herder prediction model was developed using the intensity of FDG uptake (absent, faint, moderate or intense) from the uncorrected PET images. However, as far as we are aware, lung cancer MDMs routinely assess the intensity of FDG uptake from the corrected images which is not in accordance with the original Herder model.
  The use of uncorrected images in the original Herder study (2) to distinguish between uptake categories potentially alters the perceived evidence base of the recommendations in the BTS guidelines (1) to distinguish between faint and moderate uptake according to mediastinal blood pool given that this scale of uptake was not used in the original score, has not been validated and could mean we are not using the correct category in the risk model.
  Furthermore, when considering the widespread use of the Herder score, it should be appreciated that it was formulated from patients scanned between 1997 and 2001. The Herder paper describes that “emission scans were acquired in a two-dimensional mode … and were reconstructed using ordered subset expectation maximisation with two iterations and 16 subsets followed by post smoothing… using a Hanning 0.5 filter”.
  PET cameras and reconstruction techniques have advanced considerably since then and the imaged intensity of FDG uptake within nodules <3cm in diameter acquired on modern cameras is likely to be quite different to twenty years ago. In particular it is now standard to have 3-dimensional mode acquisition and reconstructions which routinely include point spread function and time-of-flight data. This results in an increase in the SUVmax of small nodules and a potential upgrade in the avidity category of these lesions.
  A validation study by Al-Ameri et al published in 2015 (3) provides some reassurance the use of corrected images on modern scanners is still sufficiently accurate as they used corrected images to distinguish between faint, moderate and intense FDG uptake by the use of SUVmax ≤2.5, 2.6-10 and >10 as parameters respectively.
  There is now widespread use of PET-CT to characterise pulmonary nodules and increasing use of electronic databases to record patient characteristics and outcomes. We suggest that this data could be used to produce an updated version of the Herder model. For simplicity, and given its previous proven accuracy in the validating study, the same variables (age, smoking, previous history of cancer, diameter, spiculation, upper lobe) as in the original Herder model could be used. Large sample sizes and potential multi-centre data aggregation would result in model more accurate and appropriate to modern cameras than the original Herder score.
  We therefore suggest that the recommendation to use the Herder model to characterise pulmonary nodules is reviewed addressing:
  1. Clarifying whether the corrected or uncorrected PET images should be used.
  2. Determining whether the scale for distinguishing between faint and moderate uptake using mediastinal blood pool is sufficiently evidence-based.
  3. Supporting the development of an updated model using data acquired on modern day PET-CT scanner technology.

  References:
  1. Callister ME, Baldwin DR, Akram AR, Barnard S, Cane P, Draffan J, et al. British Thoracic Society guidelines for the investigation and management of pulmonary nodules. Thorax. 2015;70 Suppl 2:ii1-ii54.
  2. Herder GJ, van Tinteren H, Golding RP, Kostense PJ, Comans EF, Smit EF, et al. Clinical prediction model to characterize pulmonary nodules: validation and added value of 18F-fluorodeoxyglucose positron emission tomography. Chest. 2005;128(4):2490-6.
  3. Al-Ameri A, Malhotra P, Thygesen H, Plant PK, Vaidyanathan S, Karthik S, et al. Risk of malignancy in pulmonary nodules: A validation study of four prediction models. Lung Cancer. 2015;89(1):27-30.

  Show Less

  Conflict of Interest:
  None declared.

  • Back to top