Even the best dermatologists can’t diagnose skin cancer by eye, relying on magnifying glasses to examine suspicious blemishes and scalpels to cut tissue for analysis.
With up to more than 70 percent of biopsies coming back negative, millions of healthy patients undergo painful, costly and unnecessary procedures.
Now, using shortwave rays used in cellphones and airport security scanners, researchers at Stevens Institute of Technology have developed a technique that detects skin lesions and determines whether they are cancerous or benign – a technology that could ultimately be incorporated into a handheld device that could rapidly diagnose skin cancer without a scalpel in sight.
The work, led by Negar Tavassolian, director of the Stevens Bio-Electromagnetics Laboratory, and postdoctoral fellow Amir Mirbeik-Sabzevari, not only has the ability to reduce the number of unnecessary biopsies by 50 percent but also has the potential to disrupt a $5.3 billion diagnostic market for the most common cancer in the United States, with 9,500 Americans diagnosed with skin cancer each day.
“This could be transformative,” said first author Mirbeik-Sabzevari, whose work appears in IEEE Transactions on Medical Imaging. “No other technology has these capabilities.”
The team’s technology uses millimeter-wave radiation – the same shortwave rays used in cellphones and airport security scanners.
Millimeter-wave rays penetrate certain materials and bounce off others, which is how airport security knows if you leave your keys in your pocket as you walk through a scanner.
Just as metal reflects more energy than your body, so cancerous tumors reflect more calibrated energy than healthy skin, making it possible to identify diseased tissue by looking for reflectivity hotspots.
The latest tests were conducted on biopsies collected by surgeons from Hackensack University Medical Center.
Tavassolian and Mirbeik-Sabzevari custom built antennae to generate high-resolution images of this biopsied tissue, and found they could map the tiny tumors as accurately as lab-based testing.
Cancerous cells reflected around 40 percent more calibrated energy than healthy tissue, showing that millimeter-wave reflectivity is a reliable marker for cancerous tissue.
“We’ve shown proof-of-concept that this technology can be used for rapidly detecting skin cancer,” said Tavassolian.
“That’s a major step forward toward our ultimate goal of developing a handheld device, which would be safe to use directly on the skin for an almost instant diagnostic reading of specific kinds of skin cancer – including lethal melanomas – based on their individual reflectivity signatures.”
While the technology underpinning the device is innovative, it’s also inexpensive.
Since it involves the same basic circuitry within a cellphone, manufacturing costs would be low.
“In fact, it should be possible to keep manufacturing costs below $1,000, even at low production volumes,” said Tavassolian “That’s about the same as the magnifying tools already used by dermatologists, and an order of magnitude cheaper than laser-based imaging tools, which also tend to be slower, bulkier and less accurate than millimeter-wave scanners.”
Since millimeter-wave rays penetrate the skin, the scanners can generate real-time 3-D images of tumors that could guide surgeons and eliminate the need for multiple trial-and-error biopsies to fully remove cancerous tissue.
The devices could also be configured to interpret images automatically, and deliver basic diagnostic information – such as a warning to get checked out by a doctor – without needing a trained operator.
“We could place these devices in pharmacies, so people can get checked out and go to a doctor for a follow-up if necessary,” said Tavassolian.
“People won’t need to wait weeks to get results, and that will save lives.”
Mirbeik-Sabzevari, who began working on the technology five years ago as a graduate student in Tavassolian’s lab, is confident that this invention will prove a hit.
As a postdoctoral fellow at Stevens and the 2019 inaugural recipient of the Paul Kaplan Award for Distinguished Doctoral Research, Mirbeik-Sabzevari plans to launch a startup to commercialize the scanners.
He was the entrepreneurial lead on a customer discovery grant on this technology from the National Science Foundation.
Skin cancer is a type of cancer that arises from skin.
It is the most common form of cancer, globally accounting for at least 40% of cases [1]
It is especially common among people with light skin. Skin cancer is due to the development of abnormal cells that have the ability to invade or spread to other parts of the body.
There are three main types: basal cell cancer (BCC), squamous cell cancer (SCC) and melanoma.
The most common type is non-melanoma skin cancer, which occurs in at least 2–3 million people per year.
Of non-melanoma skin cancers, about 80% are basal cell cancers and 20% squamous cell cancers. Basal cell and squamous cell cancers rarely result in death [2].
One of the main reasons that skin cancer develops is because the DNA is damaged. DNA is the master molecule that controls and directs every cell in the body.
Damage to DNA is one of the ways that cells lose control of growth and become cancerous. DNA mutations can also be inherited.
During years few methods have been investigated in order to diagnose the skin cancer.
These methods are mainly categorized in two types which are skin biopsy, and image analysis [3].
In case of skin biopsy doctor takes a sample of skin from the suspicious area to be looked at under a microscope. Different methods can be used for a skin biopsy.
The doctor will choose one based on the size of the affected area, where it is on the patient’s body, and other factors.
In case of image analysis of damaged skin using computers, special features, like particular colours, colour variation and texture are analysed to search for the sign of cancer.
In fact, image analysis methods are based on mathematics.
For instance, Segura et al. performed a systemic analysis of melanocytic and non-melanocytic skin tumors, using dermatoscopy, RCM, and histopathology to develop a two-step method for melanoma diagnosis based on RCM features for use as an adjuvant to dermatoscopy [4].
In another extensive work Stoecker et al. analyzed asymmetry, as a critical feature in the diagnosis of malignant melanoma, using a new algorithm to find a major axis of asymmetry and calculate the degree of asymmetry of the tumor outline [5].
Thermal image analysis of damaged skin can be stated as a special type of image analysis. In this category limited works have been reported in literatures.
Flores-Sahagun et al. proposed a structured methodology for analysis and diagnosis of basal cell carcinoma (BCC) via infrared imaging temperature measurements.
They concluded that their conjugated gradients method was efficient to identify lesioned tissue (which was associated to basal cell carcinoma through a clinical exam together with skin biopsies) in all patients studied even with the use of a camera of low optical resolution (160 × 120 pixels) and thermal resolution of 0.1°C [10].
Poljak-Blazi et al. also employed infrared thermal imaging for evaluation of the tumour development and discrimination of cancer from inflammation and haematoma [11]. See also [12].
Fractals are defined to be scale-invariant (self-similar or self-affine) geometric objects.
A geometric object is called self-similar if it may be written as a union of rescaled copies of itself, with the rescaling isotropic or uniform in all directions.
Regular fractals display exact self-similarity. Random fractals display a weaker, statistical version of self-similarity or, more generally, self-affinity.
Although virtually all natural fractals are random, the concept of self-similarity is best first explored through the study of regular fractals.
The class of regular fractals includes many familiar simple objects such as line intervals, solid squares, and solid cubes, and also many irregular objects.
The scaling rules are characterized by “scaling exponents” (dimension). “Simple” regular fractals have integer scaling dimensions. Complex self-similar objects have non-integer dimension.
Therefore, it is completely incorrect to define fractals as geometric objects having “fractional” (non-integer) dimension. Fractals may be defined as geometric objects whose scaling exponent (dimension) satisfies the Szpilrajn inequality:ℵ ≥ DT(1)
where ℵ is the scaling exponent (dimension) of the object and DT is its topological dimension, i.e., Euclidean dimension of units from which the fractal object is built.
For example, in case of Brownian motion: the path of a particle, a line of dimension one, traveling for a long time over a plane region, eventually covers the entire plane, an entity of dimension two [13].
A multi-fractal system is a generalization of a fractal system in which a single exponent (the fractal dimension) is not enough to describe its dynamics; instead, a continuous spectrum of exponents (the so-called singularity spectrum) is needed.
There are limited works which have employed fractal dimension under image analysis techniques in order to analyse the skin cancer. Mastrolonardo et al. introduced the new technique of the variogram and of fractal analysis extended to the whole regions of interest of skin in order to obtain parameters able to identify the malignant lesion [14].
In another work Hall calculated fractal dimensions to represent border irregularity for early detection of melanoma [15]. In a similar work Piantanelli et al. investigated the fractal properties of skin pigmented lesion boundaries [16].
Ng and Coldman focused on using fractal concept in measuring the fuzziness of a mole. In order to overcome the problem of separation of a mole from its surrounding skin in application of variation method and the correlation method, they employed two different methods which manipulated the intensities around the border of a mole.
The first one calculated the size of the intensity surface area at different scales and the second method used the average absolute intensity difference of pixel pairs to obtain normalized fractional feature vectors [17]. See also [18].
In spite of all of these works, no work has been reported which analyses the complexity and correlation of damaged DNA through analysis of DNA walks. In this paper we use the concept of fractal dimension and the Hurst exponent in order to analyse the DNA sequences. In order to do this task we illustrate DNA walk as a random walk and by introducing the fractal dimension and Hurst exponent we compute these parameters for DNA walks which extracted from DNA sequences of healthy subjects and patients with skin cancer.
More information: Amir Mirbeik-Sabzevari et al. High-Contrast, Low-Cost, 3D Visualization of Skin Cancer Using Ultra-High-Resolution Millimeter-Wave Imaging, IEEE Transactions on Medical Imaging (2019). DOI: 10.1109/TMI.2019.2902600
Provided by Stevens Institute of Technology
