A new study uses machine learning on ultrasound images of thyroid nodules to predict the risk of malignancy

Thyroid nodules are small lumps that form within the thyroid gland and are quite common in the general population, with a prevalence as high as 67%. The great majority of thyroid nodules are not cancerous and cause no symptoms. However, there are currently limited guidelines on what to do with a nodule when the risk of cancer is uncertain. A new study from The Sidney Kimmel Cancer Center - Jefferson Health investigates whether a non-invasive method of ultrasound imaging, combined with a Google-platform machine-learning algorithm, could be used as a rapid and inexpensive first screen for thyroid cancer.

"Currently, ultrasounds can tell us if a nodule looks suspicious, and then the decision is made whether to do a needle biopsy or not," says Elizabeth Cottril, MD, an otolaryngologist at Thomas Jefferson University, and clinical leader of the study. "But fine-needle biopsies only act as a peephole, they don't tell us the whole picture. As a result, some biopsies return inconclusive results for whether or not the nodule may be malignant, or cancerous, in other words." {module In-article} 

If examining the cells of a needle biopsy proves inconclusive, the sample can be further tested via molecular diagnostics to determine the risk of malignancy. These tests look for the presence of certain mutations or molecular markers that are associated with malignant thyroid cancers. When nodules test positive for high-risk markers or mutations, the thyroid may be surgically removed. However, the standards for when to use molecular testing are still in development, and the test is not yet offered in all practice settings, especially at smaller community hospitals.

In order to improve the predictive power of the first-line diagnostic, the ultrasound, Jefferson researchers looked into machine learning or artificial intelligence models developed by Google. These applications are being used in other spaces: retail giants like Urban Outfitters use machine learning to help classify their many products, making it easier for the consumer to find an item they're interested in. Disney uses it to annotate their products based on specific characters or movies. In this case, the researchers applied a machine-learning algorithm to ultrasound images of patients' thyroid nodules to see if it could pick out distinguishing patterns. The study was published in JAMA-Oto on October 24th.

"The goal of our study was to see whether automated machine learning could use image-processing technology to predict the genetic risk of thyroid nodules, compared to molecular testing," says Kelly Daniels, a fourth-year medical student at Jefferson and first author of the study.

The researchers trained the algorithm on images from 121 patients who underwent ultrasound-guided fine needle-biopsy with subsequent molecular testing. From 134 total lesions, 43 nodules were classified as high risk and 91 were classified as low risk, based on a panel of genes used in the molecular testing. A preliminary set of images with known risk classifications was used to train the model or algorithm. From this bank of labeled images, the algorithm utilized machine-learning technology to pick out patterns associated with high and low-risk nodules, respectively. It used these patterns to form its own set of internal parameters that could be used to sort future sets of images; it essentially "trained" itself on this new task. Then the investigators tested the trained model on a different set of unlabeled images to see how closely it could classify high and low genetic risk nodules, compared to molecular test results.

"Machine learning is a low-cost and efficient tool that could help physicians arrive at a quicker decision as to how to approach an indeterminate nodule," says John Eisenbrey, PhD, associate professor of radiology and lead author of the study. "No one has used machine learning in the field of genetic risk stratification of thyroid nodule on ultrasound."

The researchers found that their algorithm performed with 97% specificity and 90% predictive positive value, meaning that 97% of patients who truly have benign nodules will have their ultrasound read as "benign" by the algorithm, and 90% of malignant or "positive" nodules are truly positive as classified by the algorithm . The high specificity is indicative of a low rate of false positives; this means that if the algorithm reads a nodule as "malignant" it is very likely to truly be malignant. The overall accuracy of the algorithm was 77.4%.

"This was such an important collaboration of surgeons and radiologists, and there's already interest from other institutions to pool our resources. The more data we feed the algorithm, the stronger and more predictive we'd expect it to become," says Dr. Cottril.

"There are so many potential applications of machine learning," says Dr. Eisenbrey. "In the future, we'd like to make use of feature extraction, which will help us identify anatomically relevant features of high-risk nodules."

Though preliminary, the study suggests that automated machine learning shows promise as an additional diagnostic tool that could improve the efficiency of thyroid cancer diagnoses. Once it becomes more robust, the approach could give doctors and patients more information in order to decide if thyroid lobe removal is necessary.

UK researchers have found no link between exposure to emissions from municipal waste incinerators (MWIs) and infant deaths or reduced fetal growth.

However, they show living closer to the incinerators themselves are associated with a very small increase in the risk of some birth defects, compared to the general population. But whether this is directly related to the incinerator or not remains unclear.

The findings come from the largest and most comprehensive analysis to date of the effects of municipal waste incinerators (MWIs) on public health in the UK.

MWIs are used to burn waste that is not recycled, composted or sent to landfill and can include materials such as paper, plastic, wood, and metal. While MWI emissions are governed by EU regulations, public concern remains around their potential impact on public health and scientific studies to date have been inconsistent or inconclusive.

The analysis, led by a team at Imperial College London and funded by Public Health England and the Scottish Government, looked at MWIs at 22 sites across the UK between 2003 and 2010.

Researchers from the UK Small Area Health Statistics Unit (SAHSU) at Imperial first analyzed concentrations of fine particles called PM10 (particulate matter measuring 10 micrometers or less in diameter) emitted from the chimneys of the incinerators as waste is burned.

Supercomputer models generated from the data showed how these particles spread over a 10 km radius around 22 MWIs in England, Scotland and Wales. The models show that MWIs added very little to the existing background levels of PM10 at ground level – with existing PM10 concentrations at ground level on average 100 to 10,000 times higher than levels emitted by the chimneys (Environment Science & Technology, 2017). {module In-article}

Using these models, the team then investigated potential links between concentrations of PM10 emitted by MWIs and an increased risk of adverse birth outcomes. In an earlier study (Environment International, 2018), they found that analysis of records covering more than one million births in England, Scotland and Wales revealed no evidence of a link between small particles emitted by the incinerators and adverse birth outcomes such as effects on birthweight, premature birth, infant death, or stillbirth, for children born within 10 km of MWIs in Great Britain.

The team’s latest findings, published in the journal Environment International, looked at the occurrence of birth defects within 10 km of a subset of 10 incinerators in England and Scotland between 2003 and 2010. In their analysis, the team used health data on more than 5000 cases of birth defects among over 200,000 births, stillbirths and terminations in England and Scotland.

They found no association between birth defects and the modeled concentrations of PM10 emitted by MWIs, but there was a small increase in the risk of two birth defects among those living closer to MWIs – specifically congenital heart defects and hypospadias (affecting the male genitalia – where the opening of the urethra is not at the top of the penis). These birth defects typically require surgery but are rarely life-threatening.

In the UK, congenital heart defects affect approximately 5.3 in 1000 births and 1.9 per 1000 males are born with hypospadias (Source: NCARDRS 2016*).

In terms of excess risk, the team estimates that the associated increase in risk for these two birth defects could be around 0.6 cases per 1,000 total births for congenital heart defects and 0.6 cases per 1,000 male births for hypospadias within 10 km of an incinerator.

Professor Paul Elliott, Director of the UK Small Area Health Statistics Unit (SAHSU) said: “Based on the available data, our findings showing that there is no significant increased risk of infant death, stillbirth, preterm birth or effects on birthweight from municipal waste incinerators are reassuring. The findings on birth defects are inconclusive, but our study design means we cannot rule out that living closer to an incinerator in itself may slightly increase the risk of some specific defects – although the reasons for this are unclear.”

Professor Mireille Toledano, Chair in Perinatal and Paediatric Environmental Epidemiology at Imperial, said: “In these studies we found a small increase in risk for children living within 10 km of an MWI being born with a heart defect, or a genital anomaly affecting boys, but did not find an association with the very low levels of particulates emitted. This increase with proximity to an incinerator may not be related directly to emissions from the MWIs. It is important to consider other potential factors such as the increased pollution from industrial traffic in the areas around MWIs or the specific population mix that lives in those areas.”

Professor Anna Hansell, Director of the Centre for Environmental Health and Sustainability at the University of Leicester, who previously led the work while at Imperial College London, added: “Taken together, this large body of work reinforces the current advice from Public Health England – that while it’s not possible to rule out all impacts on public health, modern and well-regulated incinerators are likely to have a very small, or even undetectable, impact on people living nearby.”

The team explains that while the results of the emissions studies are reassuring, they cannot rule out a link between the increased incidence of the two birth defects and the activities of the MWIs. They add that while they adjusted their results for socioeconomic and ethnic status, these may still influence birth outcomes findings. Poorer families may be living closer to MWIs due to lower housing or living costs in industrial areas, and their exposure to industrial road traffic or other pollutants may be increased.

The researchers' highlight that their findings are limited by a number of factors. Also, they did not have measurements (for the hundreds of thousands of individual births considered) of metals or chemical compounds such as polychlorinated biphenyls (PCBs) and dioxins but used PM10 concentrations as a proxy for exposure to MWI emissions – as has been used in other incinerator studies.

They add that ongoing review of evidence is needed to explore links further, as well as ongoing surveillance of incinerators in the UK to monitor any potential long-term impacts on public health.

The research was funded by Public Health England and the Scottish Government, with support from the Medical Research Council and the National Institute for Health Research.