Validation of Thermal Imaging to Detect Pneumonia

1. BACKGROUND AND SIGNIFICANCE
2. Historical Background Every day 16,000 children under the age of 5-years die. Although strides have been made in early diagnosis and treatment of pneumonia, it still accounts for about 15% of deaths of children under 5-years (1). Unfortunately care providers in low- and middle-income countries (LMICs) are limited due to lack of resources and training. Chest x-rays are especially difficult to obtain in these LMICs (14). Ironically, smartphones are readily available. Since its development in 1960, Forward-Looking Infrared (FLIR) cameras have made several technological strides. Notably they are now more affordable and compact, attaching to smartphones. FLIR cameras now have the potential to aid in early diagnosis and subsequent treatment of childhood pneumonia. FLIR imaging was first developed in 1956 with military funding at the University of Chicago. From the 1960s into the 1970s, FLIR technology was modified independently by both the Air Force and Texas Instruments Inc. for use by military planes (2). Wendell Smith first documented the use of thermal imaging in medical diagnosis in 1964. Smith required 6 minutes to take each thermograph of the horses he was studying (3). Multiple veterinary studies have been performed since then, and now thermal imaging has been validated as a means to assess musculoskeletal injuries in animals (4,5).
3. Previous Clinical Studies Supporting the Proposed Research Interestingly, very few clinical studies have been performed to validate the use of thermal imaging in medical diagnosis for human diseases. The use of thermography to diagnose lung disease was studied in Russia in 1976, 1981, and 1991 in three separate studies, but these studies were neither controlled nor randomized (6-8). In 1970 Potanin demonstrated through a case series that patients with pneumonia had increased overlying chest temperatures and patients with pulmonary embolism had decreased overlying chest temperatures (12). The most recent and impactful study evaluating pneumonia with thermal imaging was a case series performed by Wiecek et al. (2001), which showed that patients with pneumonia demonstrated a reduction in thermographic signal as they positively responded to treatment (13). More modern studies have attempted to validate infrared thermal imaging in measuring the size of brain tumors intra-operatively (9), screening for fever in children (10), and predicting length of recuperation after burn injuries (11), among others. However, there has not been a modern study validating thermal imaging in diagnosing pneumonia.
4. Rationale behind Proposed Research and Potential Benefit to Society

Pneumonia is one of the primary contributors to under-five mortality worldwide. Imaging to support the diagnosis of pneumonia can be very difficult to obtain in resource-limited settings due to lack of equipment and trained staff. Easy-to-access imaging of pneumonia with thermal imaging technology may allow practitioners to diagnose pneumonia earlier and more accurately, which would lead to decreased morbidity and mortality. In LMICs where x-ray availability is already limited in peripheral centers, excessive use of these facilities can lead to consumption of limited medical resources. Validating thermal imaging in the diagnosis of pneumonia would likely lead to cost savings, and obviation of side effects of antimicrobials and excessive radiation.

II. SPECIFIC AIMS:

1. To determine whether the FLIR ONE camera can be used to detect differences in a temperature spectrum that is clinically applicable. Temperature as measured by a temporal artery thermometer on hospitalized patients can vary from less than 96° Fahrenheit to over 104° Fahrenheit. The basis of the FLIR ONE camera and the irPNA application is to detect asymmetry of temperatures. The investigators hope to demonstrate that this technology has the capability to demonstrate this difference.
2. To validate thermal imaging with the FLIR ONE camera and irPNRA application as a sensitive modality to diagnose pneumonia. The investigators hope to demonstrate that there are few false negatives when using this modality to diagnose pneumonia.
3. To validate thermal imaging with the FLIR ONE camera and irPNRA application as a specific modality to differentiate pneumonia from other pulmonary processes, such as asthma, pulmonary embolism, and bronchiolitis. The investigators hope to demonstrate that there are few false positives when using this modality to diagnose pneumonia. III. SUBJECT SELECTION:
4. Inclusion/Exclusion Criteria Inclusion criteria are 1) patients 28-days to 17-years of age, 2) who have respiratory distress, and 3) who receive a chest x-ray. The investigators will exclude patients with known chronic lung disease.
5. Source of Samples The investigators will acquire consent from patients and parents in the Pediatric Intensive Care Unit, (PICU) Pediatric Emergency Department (ED), and Pediatric Wards at the Massachusetts General Hospital IV. SUBJECT ENROLLMENT:
6. Methods of Enrollment Trained clinical research staff will be present in the MGH PICU, pediatric wards, and ED and identify potential subjects through providers. Additionally, the EPIC electronic medical record system will be used by qualified research staff (research assistants, coordinators, managers) to electronically identify patients who may be potential candidates for the study. Subjects having a chest x-ray taken will be approached for study participation. This study will not interfere with the subject's medical care. If a patient appears to be eligible, the patient's provider will first approach the subject and ask if they are willing to be approached about a research study. If the potential subject declines, no further attempts will be made to enroll the subject. If the subject agrees to speak to someone about research, a study staff member will approach him or her about participation. Recruitment will not involve restrictions on socio-demographic factors including ethnic characteristics. Recruitment will be devoid of any procedures, which could be construed as coercive.
7. Procedures for Obtaining Informed Consent At enrollment, in a private room, the investigator/designee will obtain written informed consent from participants' parent(s) or guardian for this study. During this encounter, the investigators will remind participants that they are free to choose to take part in the research or not. The investigators will use a General Consent Form for the parents/guardians of the patients. Children 14-17 can provide assent by co-signing the General Template consent form when able. The investigators will obtain Assent from the subjects verbally, when possible.
8. Treatment Assignment and Randomization Treatments or changes in the health care plan will not be made based on this research.
9. STUDY PROCEDURES:
10. Study Parameters to be Measured In addition to obtaining thermal images with the FLIR ONE camera, The investigators will obtain several patient data points. These include 1) chief complaint, 2) temporal artery temperature, 3) cutaneous temperature, 4) ambient temperature, 5) respiratory rate, 6) oxygen saturation, 7) chest x-ray results and relevant data, 8) discharge diagnosis, and 9) disposition. Additional chest x-rays will not be obtained, the investigators will use x-rays already procured as part of the patient diagnosis and treatment approach and plan.
11. Drugs to be Used None
12. Devices to be Used FLIR ONE camera, attached to an encrypted iPhone. The investigators will also use the irPNRA application on the iPhone designed by research team member Jim Lewis to help direct the thermal imaging. The iPhone is encrypted by Partners regulations and cannot connect to a phone network or the Internet. The iPhone will remain in 'airplane mode' at all times. Biomed has already investigated and approved these devices.
13. Procedures/Surgical Interventions None
14. Data to be Collected and when the Data is to be Collected The investigators will obtain FLIR ONE thermal images within 4 hours of the subject's chest x-ray. The investigators will also collect the following data: 1) chief complaint, 2) temporal artery temperature, 3) cutaneous temperature, 4) ambient temperature, 5) respiratory rate, 6) oxygen saturation, 7) chest x-ray results and relevant data, 8) discharge diagnosis, and 9) disposition. The vitals will be obtained at the time nearest to the chest x-ray. Patient data will be collected prior to being de-identified. The investigators will use the patient name and medical record number to access their chart in the EMR. Then the following data will be obtained from the chart: 1) chief complaint, 2) temporal artery temperature, 3) cutaneous temperature, 4) ambient temperature, 5) respiratory rate, 6) oxygen saturation, 7) chest x-ray results and relevant data, 8) discharge diagnosis, and 9) disposition. The x-rays will be obtained, uploaded to the LifeImage application, de-identified, and then sent to the radiologist on our research team with the appropriate study ID. The patient information will then be de-identified and replaced with a study ID. VI. BIOSTATISTICAL ANALYSIS:
15. Specific Data Variables to be Collected The investigators will collect the following variables from the patient chart: 1) chief complaint, 2) temporal artery temperature, 3) cutaneous temperature, 4) ambient temperature, 5) respiratory rate, 6) oxygen saturation, 7) chest x-ray results and relevant data, 8) discharge diagnosis, and 9) disposition. The investigators will collect the following identifiers from the patient record prior to de-identifying them: 1) name, 2) DOB, 3) age, 4) gender, and 5) MRN. Interpretation of the x-rays will be blinded. The x-rays will be uploaded to LifeImage, de-identified, and then sent to the radiologist on our research team. The de-identified x-ray images will then be evaluated by the radiologist.
16. Study Endpoints The investigators will assess for temperature asymmetry with the FLIR ONE thermal camera with images taken of the chest. The investigators will compare these results with the results of a chest x-ray taken within the last 12 hours.
17. Statistical Methods The investigators will compare chest x-ray results to FLIR ONE thermal imaging results. The investigators will use sensitivity and specificity calculations based on the positive and negative results from those studies.
18. Power Analysis Using the standard α = 0.05, sample size = 31, observed false positive (TI+|CXR-) rate =0.38 and the observed false negative (TI-|CXR+) rate = 0.20, the power (β-1) for the test applied to the overall population was calculated at 0.26. In order to achieve a power of 0.80 with the conditions encountered in this study, The investigators would require 138 patients. VII. RISKS AND DISCOMFORT:
19. Complications of Surgical/Non-Surgical Procedures None
20. Drug Side Effects and Toxins None
21. Device Complications and Malfunctions The FLIR ONE device is a thermal camera that attaches to an iPhone. Complications of the device will not affect the patient. If it malfunctions, the investigators will fix or replace it.
22. Psychosocial (non-medical) Risks The risk of loss of confidentiality will be minimized by using study IDs on study documentation to identify subjects. The list linking the study subjects to the study IDs will be kept in a secure location, with access limited only to staff directly involved with the research study.
23. Radiation Risks There are no known radiation risks. As mentioned above, only x-rays obtained as part of the patient diagnostic and treatment plans will be used. No additional x-rays will be taken. VIII. POTENTIAL BENEFITS:
24. Potential Benefits to Participating Individuals Participants will not receive any direct benefit from their participation in the study.
25. Potential Benefits to Society Worldwide, pneumonia is one of the leading causes of mortality in children in the first 5 years of life. In both developed countries and resource poor settings, detection of bacterial pathogens by culturing blood or other specimens is time consuming, lacks sensitivity and ranges from difficult in the U.S. to impossible in the developing world. In the U.S., the inability to determine which children have a bacterial infection results in excessive use of antibiotics and increased risk of infection with multi-drug resistant organisms. In the developing world, the consequences are delays in timely initiation of antibiotic therapy and excessive referral of children for hospitalization that overwhelms referral facilities. Regardless of setting, there is an urgent need for point of care diagnostics to determine which children likely have bacterial pneumonia and will benefit from administration of antibiotics. Furthermore, aiding diagnosis of pneumonia with thermal imaging may obviate chest x-rays as the primary diagnostic tool and reduce unnecessary exposure of patients to radiation. In the PICU and the Wards, it is not uncommon that some patients require daily or even more frequent chest x-rays. If The investigators are able to validate thermal imaging as a tool to detect pneumonia, the investigators could potentially minimize the radiation exposure of these patients. IX. MONITORING AND QUALITY ASSURANCE:
26. Independent Monitoring of Source Data Given the relatively small size of this study and the lack of an investigational agent, there will not be a data and safety monitoring committee set up for the purpose of evaluating data and safety of the participants. The principal investigator, as well as the Partners IRB, shall provide oversight and review of the study protocol throughout the duration of the study.
27. Safety Monitoring There will be no official safety monitoring. The investigators anticipate the risks of participation in this research study to be minimal.
28. Outcomes Monitoring There will be no outcomes monitoring. Treatments or changes in the health care plan will not be made based on this research.
29. Adverse Event Reporting Guidelines The principal investigator will review all reports of safety or other concerns within 24 hours of being notified and communicate these to the IRB as per Partners IRB policy "Reporting Unanticipated Problems Including Adverse Events". The investigators will not require a DSMB for this study.