TAKING THE BODY’S TEMPERATURE – HISTORY AND LATEST DEVELOPMENTS

Professor Francis.J.Ring DSc MSc FIPEM FRPS ASIS, Director: Medical Imaging Research Group, School of Computing, University of Glamorgan,

(formerly Clinical Measurement, Royal National Hospital for Rheumatic Diseases Bath) on 11 July 2003

Fever was the most frequently observed condition in early medical observation. From the early days of Hippocrates, physicians have recognised the importance of a raised temperature. For centuries, this remained a subjective skill, and the concept of measuring temperature was not developed until the 16^th Century.

In modern terms we now describe heat transfer by three main modes. The first is conduction, requiring contact between the object and the sensor. The second mode of heat transfer is convection, and the third radiation. Both of the latter can now be imaged by remote detection methods.

Galilleo made his famous thermoscope from a glass tube, which functioned as an unsealed thermometer and was subject to atmospheric pressure.

Thermometry developed slowly from Galilleo’s experiments. Florentine and Venetian glassblowers in Italy made complex sealed glass containers of various shapes, to be tied onto the body surface. Temperature was assessed by the rising or falling of small beads or seeds within the fluid inside the container. Huygens, Roemer and Fahrenheit all proposed the need for a calibrated scale in the late 17^th and early 18^th century. Celsius did propose a centigrade scale based on ice and boiling water. He strangely suggested that boiling water should be zero, and melting ice 100 on his scale. It was the Danish biologist Linnaeus in 1750 who proposed the reversal of this scale, as it is known today. Although International Standards have given the term Celsius to the 0-100 scale today, strictly speaking it would be historically accurate to refer to degrees Linnaeus or Centigrade ¹.

The Clinical thermometer, which has been universally used in medicine for over 130 years was developed by Carl Wunderlich in 1868. This is essentially a maximum thermometer with a limited scale around the normal internal body temperature of 37⁰C or 98.4⁰F. Wunderlich’s treatise on body temperature in health and disease is a master-piece of painstaking work over many years. He charted the progress of all his patients daily, and sometimes two or three times during the day. His thesis, was written in German for Leipzig University and has also been translated into English in the late 19^th century².

Today, there has been a move away from glass thermometers in many countries, giving rise to more disposable thermocouple systems for routine clinical use.

Liquid crystal sensors for temperature became available in usable form in the 1960s. Originally they were painted on the skin which had previously been coated with black paint. Three of four colours became visible if the paint was at the critical temperature range for the subject. Micro-encapsulation of these substances that are primarily cholesteric esters, resulted in plastic sheet detectors. Later these sheets were mounted on a soft latex base to mould to the skin under air pressure using a cushion with a rigid clear window. Polaroid photography was then used to record the colour pattern while the sensor remained in contact. The system was re-usable and inexpensive. However, sensitivity declined over 1-2 years from manufacture, and many different pictures were required to obtain a subjective pattern of skin temperature³. The present use is for "fever" strips used to check forehead temperature on babies and young children.

Convection currents of heat emitted by the human body have been imaged by a technique called Schlieren Photography. The change in refractive index with density in the air around the body is made visible by special illumination. This method has been used to monitor heat loss in experimental subjects, especially in the design of protective clothing for people working in extreme physical environments.

Heat transfer by radiation is of great value in medicine. The human body surface requires variable degrees of heat exchange with the environment as part of the normal thermo-regulatory process. Most of this heat transfer occurs in the infra red, which can be imaged by electronic thermal imaging. Infra Red radiation was undefined before 1800 when Sir William Herschel performed his famous experiment to measure heat beyond the visible spectrum⁴. Nearly 200 years before Italian observers had noted the presence of reflected heat. John Della Porta in 1698 observed that when a candle was lit and placed before a large silver bowl in church, that he could sense the heat on his face. When he altered the positions of the candle, bowl and his face, the heat was no longer experienced.

William Herschel, in a series of careful experiments showed that not only was there a "dark heat" present, but that heat itself behaved like light, it could be reflected and refracted under the right conditions. William’s only son John Herschel repeated some experiments after his father’s death, and successfully made an image using solar radiation. This he called a "thermogram" a term still in use today to describe an image made by thermal radiation. John Herschel’s thermogram was made by focussing solar radiation with a lens onto to a suspension of carbon particles in alcohol. This process is known as evaporography⁵.

A major development came in the early 1940s with the first electronic sensor for infra red radiation. This was made from indium antimonide, and was mounted at the base of a small Dewar vessel to allow cooling with liquid nitrogen. The first medical images taken with a British prototype system, the "Pyroscan", were made at The Middlesex Hospital in London, and The Royal National Hospital for Rheumatic Diseases in Bath in 1959-1961. By modern standards these thermograms were very crude. A mark 2 Pyroscan was made for medical use in 1962, with improved images. However, the mechanical scanning was slow and each image needed from 2-5 mins. to record, line by line on electro-sensitive paper. Earlier work by the American physiologist J Hardy had shown that the human skin regardless of colour is a highly efficient radiator with an emissivity close to that of a perfect black body – 0.98. Cancer detection was a high priority subject, and with hopes that this new technique would be a tool for screening breast cancer many centres across Europe the USA and Japan became involved. A Bath surgeon Mr K.Lloyd Williams showed that many tumours are hot, and the hotter the tumour the worse the prognosis. By this time the images were displayed on a cathode ray screen in black and white. Image processing by computer had not arrived, so much discussion was given to schemes to subjectively score the images, and to look for hot spots and asymmetry of temperature in the breast. The use of false colour thermograms was only possible by photography at this time. A series of bright isotherms were manually ranged across the temperature span of the image, each being exposed through a different colour filter, and superimposed on a single frame of film.

By the mid-1970s the first computer systems had arrived. In Bath, a special system for nuclear medicine made in Sweden was adapted for thermal imaging. A colour screen was provided to display the digitised image. The processor was a pDp8, and the programme was loaded every day from paper-tape. With computerisation many problems began to be resolved. The images were archived in digital form, standard regions of interest could be selected, and temperature measurements obtained from the images. In 1974 a thermal index was devised in Bath to provide a simplified measure of inflammation. A normal range of values was established for peripheral joints, with raised values obtained in osteoarthritic joints and even higher values in Rheumatoid Arthritis. A series of clinical trials with new drugs for arthritis were performed at the RNHRD using the index to document the course of treatment.

Improvements in thermal imaging cameras have had a major impact, both on image quality and speed of image capture. Early single element detectors were dependent on optical mechanical scanning. Image resolution, spatial and thermal, were inversely dependent on scanning speed. The Bofors (Sweden) and some American imagers scanned at 1-4 frames per second. AGA (Sweden) cameras were faster at 16 frames per second, and used interlacing to smooth the image. Multiple element detector arrays were developed in the UK and were employed in cameras made by EMI and Rank. Alignment of the elements was critical, and a poorly aligned array produced characteristic banding in the image. The first significant detector for faster high resolution images was produced in Malvern by Prof. Elliott, which subsequently became known as the SPRITE detector, representing Signal Processing In The Element. This detector was used in the Rank Taylor Hobson High Resolution system called Talytherm. This camera also had a high specification infra red zoom lens, with a macro attachment. In Bath we were able to record superb images of sweat pore function, eyes with contact lenses, and skin pathology with this system some twenty years ahead of its time.

From the multi-element arrays, we now have the focal plane array detectors, with increasing numbers of pixel/elements, yielding high resolution at video frame rates. These detectors consist of a matrix of elements that are electronically scanned to produce an instant image. Good software with image enhancement and analysis is now routine in thermal imaging. (see www.medimaging.org).As standardisation of image capture and analysis become more widely accepted, the ability to manage the images, and if necessary transmit them over an intranet or internet for communication become paramount ⁶. Future developments will enable the operator of thermal imaging to use on-line reference images and reference data as a diagnostic aid. The recent SARS outbreak mainly in the Far East has resulted in rapid sales of thermal cameras to detect passengers at airports who may have a raised temperature. Although striking claims for this application are made it is unlikely to be a reliable technique without more stringent environmental conditions, and suitable preparation of every person to be imaged. This is clearly impractical in a situation where hundreds of passengers are passing through an airport every hour. They also assume that passengers in a warm climate are always cool and collected at the airport – quite an assumption!

Monitoring human body temperature is of continuing value in medicine, the technology to do this is increasingly sophisticated. One of Benjamin Franklin’s experiments in the 18^th century sent many men and ships to chart the Atlantic Ocean Gulf Stream by dipping thermometers into the sea at regular intervals - both fundamental and laborious. Today we have satellite infra red images of the same area, with remarkable ease and accuracy.

While we appreciate the modern technology with all its benefits, we can also recognise the painstaking work of the historical pioneers such as Wunderlich and Franklin.

Francis Ring

1. Ring EFJ. The History of Thermal Imaging in The Thermal Image in Medicine & Biology ed. K.Ammer & EFJ Ring p13-20. Uhlen verlag, Vienna 1995

2.Wunderlich C.A. On the Temperature in Diseases, A manual of Medical Thermometry. (translation form the German by W.Bathurst Woodman, The New Sydenham Society) London 1871

3.Flesch U. Thermographic Techniques with Liquid Crystals in Medicine in Recent Advances in Medical Thermology ed. EFJ Ring B Phillips p283-299 Plenum New York. 1984

4.Houdas & Ring Human Body Temperature, its Measurement and Regulation. Plenum New York. London 1982

5.Ring EFJ, The Discovery of Infra red Radiation in 1800. Imaging Science Journal 48: 1-8. 2000

6. Ring EFJ, Ammer K The technique of Infra red Imaging in Medicine Thermology International 10.1.7-14. 2000

Some of the answers to questions were interesting:

• 37⁰C (98.4⁰F) is the value measured by the usual method. Many biological systems operate close to this value and have a maximum only a few degrees higher; skin is damaged at 40⁰C and burns at 42⁰.
• The ear thermometers now replacing the glass version are only accurate if they use a bridge principle to compensate for ambient temperature differences. The cheap ones do not do this.
• Herschel must have used a prism made of ‘chandelier drop’ glass rather than common window glass, because the latter is only poorly transparent to Infra-Red wavelengths.

Donald Lovell