The effect of patient anxiety and depression on motion during myocardial perfusion SPECT imaging

Two psychometric instruments were administered to 275 consecutive patients (172 men, 103 women) that underwent MPI in our department within a six-month period. The patients were required to fill-in two questionnaires; the one addressed the level of state and trait anxiety of the patients prior to the procedure and the other the level of depression. Forty-nine patients (17.8 %) did not complete the questionnaires while 43 patients (15.6 %) completed them partially. It is noteworthy that the majority (63 %) of these 92 patients were women, while the majority (62.5 %) of patients who were asked to fill-in the questionnaires were men. All of these 92 patients were excluded from the study (Fig. 1).

One hundred and eighty-three (183) patients, 45 women and 138 men, accurately completed and signed the two questionnaires and their responses were evaluated in this study. Demographic information including age, years of education, occupation, and marital status, and anthropometric variables (weight and height) were also gathered. The body mass index (BMI) was calculated according to the relationship: BMI = weight in Kg/(height in m)². The difference in sample sizes between male and female patients was due to a) the smaller proportion of women who underwent MPI (37.5 %, i.e. 103 out of 275 patients), and b) the higher proportion of women (56.3 %, 58/103) compared to men (19.7 %, 34/172), who either refused to complete the questionnaires or completed them partially.

In preparation for the MPI test, information about the patient’s cardiovascular and medical history was gathered. Variables evaluated were the presence of known or suspected coronary artery disease, the presence of left ventricular systolic dysfunction (ejection fraction (EF) ≥50 % or EF < 50 %), as well as other comorbidity factors, such as cancer, serious hematologic disorders, hemodialysis, cerebrovascular disease, complicated long standing diabetes mellitus, severe connective tissue disorders, etc.

Patients were provided with the first questionnaire and were informed about the succeeding second one, by two physicians of our department, upon arrival. Both physicians had quite similar approach to the patients, in order to minimize interferences.

The Spielberger’s State-Trait Anxiety Inventory (STAI) [31, 32], is a two-page questionnaire based on a 4-point Likert scale consisting of 40 simple questions on a self-report basis. Theoretically, a 6^th grade graduate could readily comprehend and complete them in about 10 min. The STAI measures two types of anxiety: the state anxiety or temporary anxiety about an event, and the trait anxiety or daily anxiety level as a personal characteristic [31, 32]. Namely, the part of state anxiety inventory measures the extent of heightened emotions and overreaction to a situation- a perceived stressor- such as undergoing an MPI imaging. Everyone experience some anxiety, something similar to fear, which is out of proportion to the actual risk involved. The common symptoms of anxiety and fear may include uneasiness, spells of perspiration, bouts of frequent urination and muscular tension, whereas all symptoms can be associated with the release of specific neurohormonal mediators [22, 23].

The part of trait anxiety inventory measures the level of “neuroticism” and unfocused worry which stems from the personality of an individual and induces his stable tendency to respond with anxiety in the anticipation of situations [23, 24]. Higher scores on STAI questionnaire are positively correlated with higher levels of anxiety [31, 32]. Normal values for state anxiety in the Spielberger’s healthy population aged 30–59 years, was 38.1 ± 10.1 for females and 37.3 ± 9.8 for males [31]. This questionnaire had been adapted by A. Liakos and S. Giannitsis in the 1980s to suit the Greek population [33], and since, its validity had been re-evaluated by using a random sample of Greek university students and civil servants, with a mean age of 31.71, consisting of 64.8 % females. Mean anxiety state score was 43.21, mean anxiety trait score 42.79 and the resulting STAI score was 86.1 [34].

To maximize time management, the patients were provided with the second questionnaire, known as Beck Depression Inventory (BDI) [35, 36], while waiting for the delay phase of MPI. BDI is a commonly used instrument for quantifying levels of depression. It contains a 21 item self-report, each with four possible choices scored on a scale of 0 to 3. Higher total scores indicate more severe depressive symptoms. The standard cutoffs used are [35, 36]: 0–13: minimal depression; 14–19: mild depression; 20–28: moderate depression; and 29–63: severe depression.

All patients underwent ergometric (treadmill) or pharmacological (adenosine or dipyridamole) stress. SPECT imaging was performed at 10 min (stress) and 4 h (delay) following intravenous injection of 3 mCi Tl-201. A single-head GE Millennium camera, with a LEGP collimator was used for all cases. Energy window was set at 20 % at 72KeV and 30 % at 167KeV. 32 views of 40 s each, stored in a 64x64 matrix, were acquired for a 180° orbit and a total acquisition time of ~22 min. The pixel size was 6.4 mm. The raw re-projection data was reconstructed by using a filtered backprojection (FBP) algorithm (Butterworth; cutoff frequency, 0.39; power, 10) and an Ordered-Subset Expectation Maximization (OSEM) iterative reconstruction algorithm, on the Xeleris workstation (GE Healthcare).

The analysis of images for patient’s motion was based on inspection of the raw data on cinematic display, which was performed by an expert nuclear physician. The absence of motion had to be confirmed on the sinogram image as well [10]. However, motion was regarded as present only when the observer was certain of its existence. Each study was assessed for the presence of visually detectable motion, the type of motion and the grade of motion. The types of patient motion were classified as suggested in the literature [3, 5] in a) bounces (brief up or down movements, observed in <3 frames before returning to the original y-coordinate), b) shifts (up or down movement involving all the remaining frames), c) complex motion (multiple bounces or combination of bounces and shifts), d) lateral movements and translations (rotations of the body around its axis). Patient motion was scored on a scale from 1–3 as absent (0), mild (1), moderate (2) and severe (3) depending on the degree and recurrence of motion during rotating cinematic display of raw data. A single subtle event of motion throughout the dataset was identified as mild, while several single events (complex motion) were characterized as severe motion. In case of ambiguity about the largeness of a motion event, distance between lower edge of the image and lowest part of the heart silhouette was compared on the selected frames. Single events of motion, higher than 3 pixels, were characterized as severe. The aforementioned assessment of patient motion was scored for the two sets of raw data, for those of the stress phase called “Patient Motion Stress” and for those of the delay phase called “Patient Motion Delay” (Fig. 1).

The visual inspection and assessment of motion by one expert was evaluated by repeating the process almost one year after the previous one. In case of a major discrepancy, compared to the previous patient motion evaluation results (patient motion absent instead of present or vice-versa, or a motion score difference of more than 2 points regarding a specific stress or delay phase), the decision was made in cooperation with another nuclear physician expert.

A representative case of “Patient Motion” evaluation and of the corresponding “Motion Artifact” are demonstrated in Fig. 2a and b respectively.

Our final database included both qualitative (ordered and unordered) and quantitative (discrete and continuous) types of data. Qualitative variables included gender, occupation, marital status, major comorbidity, SPECT evaluation, cardiac disease class, patient motion stress, and patient motion delay. Quantitative variables included education, age, weight, height, BMI, time of cardiac disease diagnosis, state anxiety score, trait anxiety score, STAI score (sum of the two anxiety scores), and BDI score. The characteristics and the descriptive statistics (mean values and standard deviations) of the study are listed in Table 1.

Our primary interest was the investigation of possible correlations between patient mood and patient motion in the stress and delay phases of MPI. Hence, the Pearson’s coefficients of association were determined for pairs of variables.

Associations between categorical variables were also studied by generating 2x2 contingency tables and applying Fisher’s two-tailed exact test. The level of significance for all tests was set to 5 %. In the investigation of possible associations, linear or non, between pairs of the variables listed in Table 1, the two variables of motion (Patient Motion Stress & Patient Motion Delay) were considered as responses or dependent variables, while all others were considered as predictors or independent variables.

Motion quantification was not performed in our study. However, the careful visual review of projection images is considered the best way to identify any presence of motion [3, 5]. Moreover, the interobserver reliability of more than 90 % in literature [4] and the intraobserver reliability of more than 96 % in our study are in accordance with the high visual motion assessment reproducibility. Visual inspection is a widely acceptable method for detecting vertical body displacements (parallel to the axis of gantry rotation) even in the order of 3.25 mm (corresponding to half pixel in our study) rather than lateral body displacements (perpendicular to the axis of rotation and detected as a fraction of the actual motion in the particular projection). Several reports suggest that vertical motion is the most common type of motion with greater likelihood in producing clinically important artifacts [5, 6, 16].

Another limitation of our study is the acquisition of the projected images by a single-head camera. Multidetector cameras provide a significant advantage over single-detector systems due to decreased image acquisition times (~15 min compared to ~22 min in our study) and, subsequently, a lower risk of patient movement. However, it should be noted that, in case of motion the effect may be compounded, with a single motion being introduced into the dataset 2 or 3 times. Matsumoto and Germano suggest that the exposure to motion for dual gantries should be neutralized to some degree by increasing the time required for single detector acquisition [13, 14].

An issue was the unexpectedly high amount of time required to fill out the psychometric questionnaires. Another issue was the low level of education in women (31.1 %, 14/45), who were primarily from the agricultural region of the country. It could be possible that these women checked “basic education” in the questionnaire, without even having completed the primary education (primary education became compulsory in ‘30s but the law was strictly applied only in ‘70s, when secondary education became compulsory). In that respect, the educational level and the psycho-social background could be the potential factors affecting the high proportion of women who either refused to complete or inaccurately completed the questionnaires. These elderly women might experience a greater insecurity and low self-confidence, hesitating to fill out a questionnaire, since they might be afraid of potential criticism in the event they describe their condition in the questionnaire. Unfortunately, the mode of questionnaire administration, completed by the patient himself as opposed to an interview, could have effect on data quality [59]. So, it may be possible that the routine completion of such questionnaires may interfere with the smooth clinical process and cause substantial delays. An alternative type of simpler questionnaires could be more appropriate for use during clinical practice in an imaging department, which could be distributed and filled out faster and easier.