Breast Thermography

---

History

Discovery of Infrared

Sir William Herschel

In 1800, Sir William Herschel conducted an experiment that led to the discovery of infrared radiation. Sir William Herschel was the Royal Astronomer in the court of King George III. He was looking for an optical fiber material that would diminish the brightness of the sun in telescopes, experimenting with different colors of glass when he noticed that the different materials passed heat differently. Some materials did not conduct solar heat whatsoever while others made it very intense. To investigate this effect, Herschel performed a variation of Newton’s prism experiment. He used a prism to create a spectrum of light, and then he took a mercury thermometer with a blackened end and used it to measure the temperature in the different bands of light. He discovered that the temperature read by the thermometer continued to increase even after he moved the thermometer out of the colored bands of light, specifically once he moved beyond the red tinted end of the spectrum. He dubbed this area of invisible heat energy, which we now know as infrared energy, as the “thermometrical spectrum”.

Timeline of the Thermal Camera

1929 - Following World War I, Kálmán Tihanyi invented a thermal imaging device called the Evaporograph for use in British anti-aircraft defense. The Evaporograph used a layer of oil on a thin membrane to capture thermal images. A germanium lens focuses infrared radiation onto a thin nitrocellulose membrane soaked in oil. The oil will heat and diffuse in response to the focusing of the infrared radiation. The oil film will then change thickness depending on the relative temperature of the area. When a white light is shined through the membrane, the differences in thickness of the oil layer cause a color distribution across the oil film which can be photographed and interpreted for thermal information.

1947 - The US military and Texas Instruments pair up and build the first infrared line scanner, which was mounted to cargo planes and bombers during the Cold War.

1970 - Philips and EEV produce the Pyroelectric Vidicon Tube, which later became the first naval thermal imager produced by the Royal Navy.

1978 - Raytheon’s research and development team patented ferro-electric infrared detectors using barium strontium titanate.

1980 - The microbolometer is developed.

1987 - A thermal camera is used to film Predator.

2006 - Thermal cameras become commercially available to private citizens.

Timeline of Medical Thermography

300 BCE - Hippocrates observes that mud slathered onto a patient with a fever dries at different rates depending on the temperature of the skin.

1595 - Galileo Galilei invents the thermoscope.

1868 - Carl Reinhold August Wunderlich publishes Das Verhalten der Eigenwarme in Krankheiten, in which he states that normal human body temperature is 37.0 Co, also denoting that an abnormal temperature is an indicator of illness.

Late 1940s - Intra-articular temperature measurements are used to monitor steroid therapy.

1959 - The Pyroscan is first used to image the heat signature of arthritic joints.

1963 - R. Bowling Barnes demonstrates that a thermogram may be utilized for medical diagnoses, specifically in terms of vasculature.

1960s-1970s - Oscilloscope displays become a part of thermal imaging and multiple exposure photography allows for the first color thermogram to be made.

1978 - Thompson et al. publish findings about tumor detection using thermography.

1984 - Fushimi et al. demonstrates use of thermography to classify autonomic neuropathy in diabetes mellitus.

Dr. Ray Newton Lawson and the Conception of Breast Thermography

Dr. Ray Newton Lawson was a Canadian surgeon who pioneered the idea of using thermal imaging to diagnose breast cancer. Dr. Lawson was born in London, Ontario in 1912. He received his undergraduate medical degree in 1937 following his graduation from the University of Western Ontario. He worked as a doctor in the Royal Canadian Air Force during World War II. In 1944 he became a Edward W Archibald Residential Fellow in Experimental Surgery at McGill University in Montreal, where he studied surgery. Following his graduation from the program, he joined the staff at the Royal Victoria Hospital in Montreal. While in his practice, Dr. Lawson frequently performed surgery on cancer patients, and he grew interested in breast cancer. In the 1950s, he observed that tumors were on average 1.75-3.5 degrees warmer than surrounding tissue. This led him to hypothesize that one could use heat detection to detect the presence of tumors. Since heat detection does not require contact with the body, this technology presented an attractive alternative to traditional tissue biopsies. In 1956, Lawson read about the use of infrared imaging using the Baird Evapograph, which had been used by the American Department of Defense for surveillance and heat-seeking missile development. After getting into contact with Dr. P Baird, Lawson secured access to the device and from 1956 to 1962 he worked in conjunction with Infrared Industries and the National Cancer Institute of Canada to adapt the device to medical use. The resultant device was dubbed the Thermoscan and released for sale. The Thermoscan worked by detecting the thermal energy emitted by the subject, as opposed to thermal energy reflected by the subject as previous technologies had done. The device took images by using a mirror to reflect heat emitted from the subject onto a second parabolic mirror. The parabolic mirror then focused the heat onto an infrared detecting cell made of indium antimonide mounted on glass. Much like modern detection systems, the infrared detecting cell’s resistance changed in response to incident thermal energy. The change in resistance affects the voltage output of the device and this varying voltage is connected to a modulator tube, whose brightness fluctuated with fluctuating voltage. The modulator tube scanned over polaroid film, generating a thermographic image.

How Thermography Works

Thermal Energy

Thermal energy is part of the infrared energy spectrum between 8-15 μm wavelengths. All matter is made up of atoms in constant motion. This constant motion generates kinetic heat and releases electromagnetic radiation. This radiation can be represented as radiant flux (φ), which is measured in units of Watts. Radiant flux specifically refers to the radiant energy that exits an object. An object’s temperature is positively related to radiant flux emitted by that object. Thus radiant flux can be measured and used to estimate an object’s temperature. The rate at which an object releases radiant flux per unit area and unit wavelength at a fixed temperature is called spectral radiance. Mathematically it can be expressed as:

$I(\lambda ,T)=\frac{2h{c}^2}{{\lambda }^5}\times \frac{1}{\frac{hc}{e^{\lambda kT}}-1}$

Where I is spectral radiance, λ is wavelength, T is temperature in Kelvin, h is the Planck constant, k is the Boltzmann constant, and c is the speed of light. This equation is written in reference to an ideal blackbody, an object that absorbs all electromagnetic energy directed at it. Although a true ideal blackbody does not exist, this equation can be used to describe how an object radiates heat. Using this equation, one can identify the characteristic wavelength that corresponds to a blackbody’s temperature. This characteristic wavelength can then be converted to an absolute temperature using Wein’s Displacement Law:

${\lambda }_{max}=\frac{b}{T}$

Where λ is the aforementioned characteristic or maximum wavelength derived using the spectral radiance equation, b is Wein’s displacement constant, and T is the blackbody’s absolute temperature in Kelvin. This relationship is also used in the design of thermal cameras. Since the range of temperatures that a living object can produce is known, this equation can be used to identify the wavelength of the infrared radiation emitted from the object, which can then be used to identify what sort of sensors should be used in the camera. Camera calibration is further influenced by Kirschoff’s radiation law, which states:

${\varphi }_{incident}={\varphi }_{transmitted}+{\varphi }_{absorbed}+{\varphi }_{reflected}$

Where φ is radiant flux. In medical thermal imaging, the transmission term is eliminated as living things do not transmit radiant flux. Further, for an ideal blackbody, reflected radiant flux is also eliminated since blackbodies do not reflect any radiant flux. This results that for a blackbody, any radiant flux that is absorbed must also be emitted. Since perfect blackbodies do not exist, it follows that the maximum radiant flux emitted by an object can not exceed that emitted by a blackbody at the same temperature. This information is used to calibrate thermal imaging cameras to give more accurate temperature readings. The final equation relevant to thermal imaging is the Stefan-Boltzmann Law:

$M_b=A\sigma T^4$

Where M is the emitted spectral radiance per unit surface area, A is the surface area of the object, σ is the Stefan-Boltzmann constant, and T is the absolute temperature of the object in Kelvin. This equation expresses the amount of heat energy that can be released by the exposed surface area of a blackbody.

Emissivity

Emissivity is a physical property of a material and is a ratio of the amount of heat energy emitted by an object as compared to an ideal blackbody at the same temperature. The amount of heat energy released by an object is dependent on emissivity. The emissivity of an object is used in thermal imaging to calibrate the thermal camera. If an object has low emissivity, then it doesn’t release very much heat energy, and this can lead to an inaccurately low temperature reading from the thermal imager unless it is calibrated with the object’s emissivity value.

Thermal Camera Organization and Function

Thermal cameras are made up of a lens, a sensor, and a means of processing the heat signal such that it is converted into an image. The lens, once made of germanium and now of chalcogenide glass, focuses the heat radiation onto a sensor array. The sensor array converts the detected heat into an electric signal. The processor then interprets the signal produced by each pixel of the sensor array and correlates the signal to a color. This conversion results in a colored (or black and white) image where the colors correspond to the varying temperatures of the measured objects.

Microbolometers

Bolometer Diagram

As mentioned above, a sensor array is used to convert the thermal energy into an electrical signal. Most thermal imagers use devices called microbolometers to achieve this conversion. When microbolometers absorb thermal energy, their electrical resistance changes. These resistance changes can then be interpreted by a signal processing system and used to identify differences in temperature in an object.

Signal to Temperature Conversion

The thermal signal received by a thermal imager is not solely the heat emitted by the object being imaged, but also includes heat emitted by the surroundings. In terms of signal, this can be represented as:

$s_{object}=s_{total}-s_{reflected}-s_{atmosphere}$

Where s corresponds to the signal. The atmospheric and reflected signals are typically calculated based on information given by the user. Once the object signal has been identified, the temperature of the object can be calculated using the following equation:

$T=\frac{B}{ln(\frac{R}{s})+F}$

Where T is object temperature, B, R, and F are coefficients related to the lens and camera, and s is the signal from the object measured earlier.

Confounding Influences

Several factors in the environment can affect the quality and accuracy of a thermographic reading. Primarily these effects add or remove infrared radiation perceived by the sensor, which can then alter the calculated temperature.

Background Reflectance

Since there are no ideal blackbodies, all objects must reflect some amount of heat energy. The energy reflected by objects surrounding the object of interest is referred to as background reflectance and represents the aggregate temperature of the surrounding environment. However, this background reflectance has a small influence on the temperature measurement of biological objects, resulting in only ~5% error. Background reflectance is usually accounted for by recording the air temperature. However, this accounting is only an approximation, since background temperature is not uniform.

Atmospheric Effects

Atmospheric effects are those which are created by gasses in the atmosphere. Ozone, water vapor, and carbon dioxide all absorb infrared radiation, with water vapor being the most absorbent of the three main gasses. The effect of atmospheric absorption of infrared radiation is compounded by the distance of the object from the detector.

Distance Effects

Increasing the distance between an object and a thermal imager increases the amount of radiation lost to the atmosphere, which can impinge on the accuracy of the measurement. Further, increasing distance between object and imager also reduces the resolution of the measured temperatures.

Focus

Focus comes into play in the interpretation of a thermographic image. Reduced focus or image blurring can reduce the resolution of the image, which can make differentiating between areas with small temperature differences difficult. Further, lack of focus can cause a blending effect between the intended object and the background. This blending can cause the measured object temperature to be influenced by the background temperature. For example, if the background is warmer than the object, blurring between object and background can result in the object being recorded as having a higher temperature than it actually does.

Noise

Noise is a product of random sensitivity fluctuations in the microbolometers. Since microbolometers are in fact made up of arrays of pixels, any fluctuations in the pixel responses will cause noise to build up in the signal. This can be corrected using a non-uniform correction function, which recalibrates individual pixels to the mean pixel sensitivity, thus reducing noise.

Resolution

The resolution, or smallest resolvable distance on a single pixel of an image, of a thermographic image is dependent on the following equation:

$MSS=d\times {\theta }_{IFOV}$

Where MSS is the minimum measurement spot size, or resolution, d is the distance between the camera and the object in meters, and θ is the instantaneous field of view angle in radians. The instantaneous field of view angle is the angle over which heat is detected by the camera.

Output Image

Black and White Thermal Map of Breast Tissue

False Color Thermal Map of Breast Tissue

The output image produced by a thermal imager comes in one of two forms. The image can be black and white, where white corresponds to higher temperatures while black corresponds to colder temperatures. Images can also be displayed in false color, which refers to a color gradient that does not correspond to the actual color of objects. In false color thermographic images, warm colors such as red and yellow correspond to high temperature while cool colors such as green and blue correspond to lower temperature.

Breast Thermography

Description

According to the American Academy of Thermology, breast thermography uses measurements of skin temperature to observe changes in breast tissue. Images are analyzed both to identify changes in how the skin gives off heat and to perform thermovascular mapping.

How It Should Theoretically Work

Tumors have an increased metabolism and angiogenesis. This, combined with increased nitric oxide levels in the tumor itself, releases heat, which would increase the temperature of the tissue. This would create an asymmetric heat profile between a cancerous and noncancerous breast. The aim of thermographic breast screening is to detect this asymmetry.

Advantages and Disadvantages

Advantages

• Non invasive
• Low cost
• Can be used to produce a large volume of screenings quickly

Disadvantages

• Low sensitivity
• Low specificity
• Inconsistent accuracy even across multiple studies
• Diagnosis is dependent on human image interpretation

Comparison to Mammography

According to the FDA, breast thermography is not a substitute for mammography, but rather has been approved to be used as an “adjunctive” tool to be used in concert with mammography. Thermography does have some advantages compared to mammography, despite its lower accuracy in detection. Mammography is not usable for patients with dense breasts or young patients with small breasts. Further, mammography requires expensive equipment and trained medical professionals that may not be available in developing countries. Thermography has additional allure in that it is noninvasive and does not use ionizing radiation to derive the image, unlike mammography, which can cause physical discomfort and uses X-rays to generate the image.

Directions for Improvement

One of the primary limitations on the use of thermography is the need for trained professionals to interpret the thermographic images. To this end, there has been a shift to use machine learning to train computers to interpret thermographic images. This shift aims to improve the screening accuracy of thermography by having computers analyze images to detect asymmetry in breast heat distribution as opposed to using human technicians. This change does face some challenges due to the images available for use. The training of any artificial system is dependent on having access to images that can be used to train the algorithm. Currently the images available to the public are inconsistent in their quality and acquisition procedures and sparse in their range, so while this shift shows promise, there is still much room for improvement.

This article "Breast Thermography" is from Wikipedia. The list of its authors can be seen in its historical and/or the page Edithistory:Breast Thermography. Articles copied from Draft Namespace on Wikipedia could be seen on the Draft Namespace of Wikipedia and not main one.

---