John E. Brimm

Estimation of the volume of lung below the left atrium using computed tomography

Gravity has significant effects on the microvascular pressure in the lungs and thus on regional fluid filtration in the lungs. Below the level of the heart, gravity increases the microvascular pressure with respect to the left atrial pressure; above the level of the heart, microvascular pressure is less than atrial pressure. To assess the contribution of gravity to fluid filtration in the lungs independent of left-heart filling pressure, the distribution of lung volume above and below the left atrium must first be determined.To permit calculation of the contribution of hydrostatic pressure per unit of lung volume, 26 computed tomographic scans of the entire chest were traced and measured, marking the level of the center of the left atrium (LAL) on each slice. The intrathoracic volume above and below the left atrium was measured by multiplying scan slice thickness by the area of the lung above or below the LAL measured on each scan slice. On 16 scans, intrathoracic volumes of 1-cm horizontal layers of lung above and below the LAL were also calculated by measuring corresponding segments of the transverse scan slices. The calculations indicate that in the supine posture more lung is dependent than when upright, and that elevation of a patient to 30° reduces the volume of lung below the LAL nearly as much as does the upright posture.

Automated documentation and analysis of burn size

The digital recording and processing of information on burn size provides a useful adjunct to the care of patients with burn injuries in the critical care setting. Record keeping is improved and accuracy is enhanced using a simple, portable system based on the widely available IBM Personal Computer. Diagramming burns on a graphic outline of a human body with color coding of burn depth produces a visual representation of the burned patient. Computations of areas of burn are automatically produced from the graphic images.

Clinical Basis and Use of an Automated ICU Testing System

Tools for monitoring respiration to prevent accidents and to detect catastrophic failures are still in their infancy. These tools are essential, however, because the relative danger of accidental airway disconnection or ventilator malfunction increases directly with a patient’s dependence on ventilator support. The most dependent patients have drug or disease induced muscle paralysis and cannot develop spontaneous ventilatory force. At the other extreme, patients on intermittent mandatory ventilation (IMV) require only a small number of breaths per minute to supplement their own ventilatory effort. A continuous monitor of expiratory volume, respiratory rate, and airway pressure is needed to prevent catastrophic failures. Unfortunately, this full mix has rarely been available, principally because of the unreliability of airflow sensors .

Computerized ICU Data Management

Much effort has been spent in developing ICU computer systems to provide crisis alarms for events such as cardiac arrest and disconnection from a respirator. With the exception of arrhythmia detection in CCU care, however, alarms based on vital signs have not had significant success. This is in large part because alarms from a central computer duplicate those from bedside hardware.

Instrumentation and Automation of an ICU Respiratory Testing System

Since 1971, a commercially available ICU minicomputer system has been used as the basis for developing an intermittent respiratory testing system. Airway flow and pressure signals and instantaneous gas concentrations, measured by a mobile mass spectrometer, are used to perform a variety of respiratory function tests at the ICU bedside. The determinations include respiratory work and mechanics, O2 consumption and CO2 production, intrapulmonary shunt fraction, and spirometry. A telephone communications facility extends use of the computer to remote hospitals. Data analysis algorithms support the use of intermittent mandatory ventilation (IMV) and continuous positive airway pressure (CPAP).

Accuracy of radiographic lung volume using new equations derived from computed tomography.

This study compared radiographic measurements of regional lung volume with corresponding physiologic measurements. The database included 48 normal, supine adults who had previously undergone spirometry and helium-dilution lung volume measurements to determine physiologic functional residual capacity (G-FRC) and total lung capacity (G-TLC). Chest x-rays had also been obtained for these subjects at functional residual capacity and total lung capacity.To calculate radiographic lung volumes at functional residual capacity and total lung capacity (R-FRC and R-TLC, respectively), we traced the lung outline from the chest x-ray, and digitized each outline. We then calculated total and regional lung volumes, using equations previously derived from computed tomographic scans. R-FRC and R-TLC were closely correlated (r = 0.973, maximum SEE 6.8%) with G-FRC and G-TLC, whether or not separate equations were used for regions below the diaphragm. Using only maximum anteroposterior diameter reduced the r-value to 0.957 (SEE 9.6%). The accuracy of our method of radiographic measurement of regional lung volume was validated by the concurrence between the physiologic and radiographic measurements.

Measuring Lung Volumes from Chest Films Using Equations Derived from Computed Tomography

Computed tomographic scans show true cross-sectional area of a segment of the chest. Measurement of the cross-sectional area of the lung at several adjacent levels permits calculation of a geometrically defined volume. CT scans from 26 cases were used to derive equations to predict regional volume from measurements which can be obtained from plain PA and lateral chest films. Separate equations were derived for slices above and below the top of the diaphragm. The best correlation between linear dimensions and true volume was obtained with equations that used lung width and antero-posterior (AP) diameter of each scan, maximum AP lung diameter, and relative scan level (apex to base). These equations predicted the volumes of individual slices above the diaphragm with a correlation coefficient (r) of 0.99 on the right, 0.97 on the left. Below top of the diaphragm, r was 0.91 on the right, 0.92 on the left. Total lung volume was predicted with an r of 0.98 (s.e. 4.8%) on the right and 0.97 (s.e. 5.3%) on the left. Using total chest width instead of AP diameter of each slice reduced r to 0.96 for the volume of either lung. This method compares favorably with previous regression or geometric approximation methods for total lung volume and also makes it possible to obtain estimates of portions of lung volume from chest films.

Extracting Information from Blood Gas Analyses

Incorrect decisions regarding patient management can result when physicians misinterpret blood gas analyses because they do not calculate, or do not know how to calculate, important derived values. To assure complete analysis of available data from the patient, we have developed a computer-based system for extracting maximum information from the tests performed. Whenever a patient’s arterial and/or venous blood gas data are entered, the system corrects them to body temperature, calculates the O2 saturation, base excess, O2 content and predicted PaO2 for various FIO2’s, and provides an interpretation of acid-base status. It then integrates the blood gas data with other data collected by our computerized cardiopulmonary testing system. The system automatically searches for time-windowed results of various respiratory and hemodynamic tests—cardiac output determinations, end-tidal gas concentrations and gas exchange. Then it computes the available oxygen, shunt fraction, arterial-alveolar oxygen difference, dead space-to-tidal volume ratio, O2 consumption, and other values.

Computer Estimation and Monitoring of Nutrition Programs for Burned Patients

A computer is used to estimate and monitor nutritional requirements of significantly burned patients. Established nutritional support formulas are programmed in the computer, as are the components of each of the nutritional solutions used for fluid resuscitation, enteral feedings, and parenteral nut

Computer Based Fluid Balance Records

Fluid balance data is essential for decision making regarding cardiovascular, respiratory, renal, and metabolic function; however, manual charting of fluid balance is characterized by three principal deficiencies: it is inaccurate, it reflects fluid volumes but not their compositions; and it masks long-term effects. In order to assess the accuracy of manual charting, we studied the records of 30 randomly selected patients, whose ICU admissions lasted four or more days, from five 6-bed ICU’s in our hospital. We found that 30 percent of the daily records were incorrect, and that only four patients had fully error-free records. Most errors were arithmetical; however, errors persisted even when calculators were used. Frequently, hourly entries were omitted from totals, and intakes were confused with outputs. In one case, a net intake of two liters was confused with a net output. These mistakes, as expected, were correlated with the number of intakes and outputs, indicating that patients with the most complex fluid management problems are least likely to have accurate records. The implications of this inaccuracy are clear.