Method for improving clinical data quality in positron emission tomography

• US 20050129170A1
• Filed: 11/12/2004
• Published: 06/16/2005
• Est. Priority Date: 11/14/2003
• Status: Active Grant

• Current Assignee: Siemens Medical Solutions USA Incorporated (Siemens AG)
• Original Assignee: CTI PET Systems, Inc. (Siemens AG)
• Inventors: Casey, Michael E., Bendriem, Bernard, Watson, Charles
• Granted Patent: US 7,649,176 B2
• Time in Patent Office: Days
• US Class Current: 378/5
• CPC Class Codes: G01T 1/2985 In depth localisation, e.g. using positron emitters; Tomographic imaging (longitudinal and transverse section imaging; apparatus for radiation diagnosis sequentially in different planes, steroscopic radiation diagnosis); using external radiation sources A61B6/02
• CPC Class Codes: G01T 7/005 calibration techniques stabilization of spectrometer G01T1/40

Abstract

A method for improving clinical data quality in Positron Emission Tomography (PET). The method provides for the processing of PET data to accurately and efficiently determine a data signal-to-noise ratio (SNR) corresponding to each individual clinical patient scan, as a function of a singles rate in a PET scanner. The method relates an injected dose to the singles rate to determine SNR(D_inj), and provides an accurate estimate of a quantity proportional to SNR, similar in function to the SNR(D_inj). Knowledge of SNR(D_inj) permits determination of peak SNR, optimal dose, SNR deficit, dose deficit, and differential dose benefit. The patient dose is fractionated, with a small calibration dose given initially. After a short uptake, the patient is pre-scanned to determine T, S, and R. An optimal dose is then determined and the remainder injected.

Claims

1. A method for improving data quality in Positron Emission Tomography (PET), said method comprising the steps of:

(a) determining a generic functional form T(p) for a true count rate response in a PET scanner, wherein p is a selected parameter;

(b) determining a generic functional form R(p) for a randoms count rate response in the PET scanner;

(c) determining a value of a true count rate T in a specific clinical patient scan;

(d) determining a value of a randoms count rate R in the specific clinical patient scan;

(e) determining a value of a selected parameter p in the specific clinical patient scan;

(f) calibrating said generic functional form of T(p) using said values of T and p to acquire a calibrated T_pat(p), whereby said calibrated T_pat(p) accurately represents an actual count rate response for the patient and scan configuration; and

(g) calibrating said generic functional form of R(p) using said values of R and p to acquire a calibrated R_pat(p), whereby said calibrated R_pat(p) accurately represents an actual count rate response for the patient and scan configuration.

2. The method of claim 1 further comprising the step of:

(h) determining data quality metrics relative to the patient and scan configuration, said step of determining data quality metrics being accomplished using at least one of said T_pat(p) and said R_pat(p).

3. The method of claim 1 wherein p is a directly acquired quantity used to provide a count loss correction in a PET tomograph.

4. The method of claim 1 wherein p is a singles event rate in a PET tomograph, said singles event rate being selected from an integral singles event rate and a vector comprising an individual detector single event rates.

5. The method of claim 1 wherein p is output from device measuring detector dead time.

6. The method of claim 5 wherein said device measuring detector dead time is a live-time clock.

7. The method of claim 1 wherein said randoms rate is estimated using one of a group of methods including at least estimating said randoms rate directly from a delayeds sinogram, estimating said randoms rate as a fraction of a total delayeds rate assuming a uniform distribution and using an attenuation mask, and estimating said randoms rate from a spatial distribution of said singles rates.

8. The method of claim 1 further including the step of:

(i) determining a calibrated scatter count rate S_pat(p), said step of determining a scatter count rate S_pat(p) including the steps of;

(1) determining a value of a scatter fraction S_f in the specific clinical patient scan by;

S_f=S/(T+S); and

(2) applying said scatter fraction value to T_pat(p) according to;

S_pat(p)=└

S_f/(1−

S_f)┘

T_pat(P).

9. The method of claim 1 further comprising the step of:

(j) determining a true plus scatter count rate (T+S)(p), said step of determining a true plus scatter count rate including the steps of;

(1) determining a generic functional form (T+S)(p) for count rate response;

(2) determining a particular value for (T+S) in said specific clinical patient scan; and

(3) calibrating said generic functional form of (T+S)(p) using said particular values of (T+S) and said selected parameter p to acquire a calibrated (T+S)_pat (p), whereby said calibrated (T+S)_pat(p) accurately represents an actual count rate response for the patient and scan configuration.

10. The method of claim 9 wherein T(s) closely approximates (T+S)(s) in shape and wherein (T+S) defines a net trues rate, wherein (T+S) is measured directly, and wherein said trues rate is estimated by subtracting an estimate of a mean scatter distribution.

11. The method of claim 9 further comprising the step of:

(k) determining data quality metrics relative to the patient and scan configuration, said step of determining data quality metrics being accomplished using at least one of a scatter count rate S_pat(p) and said true plus scatter count rate (T+S)_pat(p) in conjunction with at least one of said true count rate T_pat(p) and said randoms count rate R_pat(p).

12. The method of claim 9 wherein at least one of said steps of determining a generic functional form T(p);

determining a generic functional form R(p); and

determining a generic functional form (T+S)(p) is performed using at least one measurement on a phantom to acquire T_phant(p), R_phant(p) and (T+S)_phant(p), respectively.

13. The method of claim 11 wherein T_pat(p) is computed from T_phant(p) by calibrating with T_pat(p_pat^meant) according to:

14. The method of claim 12 wherein (T+S)_pat(p) is computed from (T+S)_phant(p) by calibrating with

15. The method of claim 9 wherein said true count rate T, said randoms count rate R, a scatter count rate S, said true and scatter count rate T+S, said true count rate response function T(p), said randoms count rate response function R(p), said scatter count rate response function S(p), and said true and scatter count rate response function (T+S)(p) refer to count rates selected from integral count rates and local count rates determined from weighted sums over particular regions of interest in available data, said available data being distributed in at least one of space, energy and time.

16. The method of claim 9 further comprising the step of:

(1) repeating said steps of (f) calibrating said generic functional form of T(p) using said values of T and p to acquire a calibrated T_pat(p);

(g) calibrating said generic functional form of R(p) using said values of R and p to acquire a calibrated R_pat(p); and

(j)(3) calibrating said generic functional form of (T+S)(p) using said particular values of (T+S) and said selected parameter p to acquire a calibrated (T+S)_pat(p) to acquire said T_pat(p), R_pat(p), and (T+S)_pat(p), respectively, corresponding to distinct regions of interest.

17. The method of claim 9 wherein said step of (f calibrating said generic functional form of T(p) using said values of T and p to acquire a calibrated T_pat(p) includes the steps of:

(1) selecting said selected parameter p to be monotonically variable with respect to activity A_pat potentially present in the patient;

(2) determining a generic functional form p(A) using at least one of at least one measurement on a phantom and theoretical considerations;

(3) inverting said p(A) to determine a generic functional form A(p);

(4) determining a value of A in a specific clinical patient scan, wherein A represents at least a quantity proportional to the actual total activity present;

(5) calibrating said generic functional form of A(p) using said value of A to acquire a calibrated A_pat(p) according to;

$A_{\mathrm{pat}}<br><br>\left(p\right)=\frac{A_{\mathrm{pat}}<br><br>\left(p_{\mathrm{pat}}^{\mathrm{meas}}\right)}{A_{\mathrm{phant}}<br><br>\left(p_{\mathrm{pat}}^{\mathrm{Meas}}\right)}<br><br>A_{\mathrm{phant}}<br><br>\left(p\right);<br><br>$

(6) inverting A_patA(p) to determine p(A_pat); and

(7) substituting p(A_pat) into T(p) to acquire T(p(A_pat));

wherein said step of (g) calibrating said generic functional form of R(p) using said values of R and p to acquire a calibrated R_pat(p), includes the step of;

(8) substituting p(A_pat) into R(p) to acquire R(p(A_pat)); and

wherein said step of (j)(3) calibrating said generic functional form of (T+S)(p) using said values of (T+S) and p to acquire a calibrated (T+S)_pat(p), includes the step of;

(9) substituting p(A_pat) into (T+S)(p) to acquire (T+S)(p(A_pat)).

18. The method of claim 17 wherein said activity A_pat is a dose D_inj of radiopharmaceutical injected in the patient, D_inj being calibrated to said generic form A(p) according to:

19. The method of claim 18 further comprising the step of correcting said injected dose D_inj for at least one of decay correction to a standard uptake period and for excretion.

20. The method of claim 11, in said step of (k) determining data quality metrics relative to the patient and scan configuration, wherein said data quality metrics include at least one of a noise equivalent count rate NECR determined according to:

21. The method of claim 11, in said step of (k) determining data quality metrics relative to the patient and scan configuration, wherein said data quality metrics include at least one of an optimal noise equivalent count rate NECR, an optimal pseudo-noise equivalent count rate PNECR, an optimal signal-to-noise ratio SNR, an optimal injected dose, a dose deficit, an SNR deficit, an NECR deficit, a PNECR deficit, and a differential dose benefit.

22. The method of claim 11 further comprising the step of developing a database of said data quality metrics relating patient characteristics including at least weight, body mass index, sex, disease, diabetes state, and anatomical position to optimal dose and peak SNR.

23. The method of claim 22 further including the step of statistically correlating said optimal dose for each scan completed into said database.

24. The method of claim 22 further including the steps of:

(a) comparing said patient characteristics for a new patient with said database; and

(b) estimating said optimal dose for said new patient.

25. The method of claim 1, before said step of (a) determining a generic functional form T(p) for a true count rate response, further comprising the steps of:

(aa) injecting a patient with a calibration dose of a radiopharmaceutical, said calibration dose being a fractionated portion of a patient dose of the radiopharmaceutical; and

(ab) pre-scanning the patient;

and after said step of (g) calibrating said generic functional form of R(p), further comprising the steps of;

(h) determining a calibrated true plus scatter count rate (T+S)_pat(p);

(i) determining an optimal dose of the radiopharmaceutical;

(j) determining a remainder dose as a difference between said optimal dose and said calibration dose; and

(k) injecting said patient with said remainder dose.

26. A method for improving data quality in Positron Emission Tomography (PET), said method including the step of:

(a) partitioning a total duration of a scan among selected regions of the body according to predetermined values of at least one of a signal-to-noise ratio (SNR), noise equivalent count rate (NECR) and pseudo-noise equivalent count rate (PNECR), wherein said SNR, said NECR, and said PNECR are each specific to each selected region of the body.

27. The method of claim 26 wherein said predetermined values are measured values.

28. The method of claim 27 wherein said measured values are determined using the steps of:

(a) pre-scanning the body over each selected region for a duration, said duration being a fraction of a total scan duration required for a patient scan, a remainder of said total scan duration being divided among said selected regions of the body;

(b) measuring a true count rate;

(c) measuring a randoms count rate;

(d) measuring a scatter count rate;

(e) determining said SNR;

(f) determining said NECR; and

(g) determining said PNECR.

29. The method of claim 26 wherein said predetermined values are predicted values.

30. The method of claim 29 wherein said predicted values are determined by correlating at least one patient characteristic with a database containing results from other patient scans relating said at least one patient characteristic to said SNR, said NECR, and said PNECR in the selected regions of the body, said at least one patient characteristic including at least injected dose, uptake period, weight, sex, and body mass index.

31. The method of claim 26 wherein at least one of said predetermined values is a measured value and wherein at least one of said predetermined values is a predicted value.

32. The method of claim 26 wherein said step of (a) partitioning a total duration of a scan among selected regions of the body includes the steps of:

(1) operating the PET tomograph in a step-and-shoot mode, wherein the PET tomograph bed is stepped through selected bed positions; and

(2) varying acquisition time of each of said selected bed positions.

33. The method of claim 26 wherein said step of (a) partitioning a total duration of a scan among selected regions of the body includes the steps of:

(1) operating the PET tomograph in a continuous bed motion mode, wherein the PET tomograph bed is continuously moving during a patient scan; and

(2) varying a bed feed rate between selected bed positions.

34. The method of claim 26 further comprising the step of:

(b) determining an optimal scan duration in each selected region of the body, whereby said SNR is equalized among each said selected region of the body.

35. A method for improving clinical data quality in Positron Emission Tomography (PET), said method comprising the steps of:

(a) injecting a patient with a calibration dose of a radiopharmaceutical, said calibration dose being a fractionated portion of a patient dose of the radiopharmaceutical;

(b) pre-scanning the patient;

(c) determining a trues rate (T_pat);

(d) determining a scattered events rate (S_pat);

(e) determining a randoms rate (R_pat);

(f) determining a singles rate

$s_{\mathrm{pat}}^{\mathrm{meas}};<br><br>$

(g) determining an optimal dose of the radiopharmaceutical using T_pat, S_pat, R_pat and

$s_{\mathrm{pat}}^{\mathrm{meas}};<br><br>$

(h) determining a remainder dose as a difference between said optimal dose and said calibration dose; and

(i) injecting said patient with said remainder dose.

36. The method of claim 35 wherein said step of determining a trues rate is performed by:

37. The method of claim 36 wherein T(s) and R(s) are measured for a given scanner at at least one value of s using a phantom.

38. The method of claim 36 wherein T(s) closely approximates (T+S)(s) in shape and wherein (T+S) defines a net trues rate, wherein (T+S) is measured directly, and wherein said trues rate is estimated by subtracting an estimate of a mean scatter distribution.

39. The method of claim 36 wherein said randoms rate is estimated using one of a group of methods including at least estimating said randoms rate directly from a delayeds sinogram, estimating said randoms rate as a fraction of a total delayeds rate assuming a uniform distribution and using an attenuation mask, and estimating said randoms rate from a spatial distribution of said singles rates.

40. The method of claim 36 further including the step of estimating a signal-to-noise ratio (SNR) as a function of said singles rate.

41. The method of claim 40 further including the step of relating said calibration dose to said singles rate by calibrating a phantom measurement of activity compared to said singles rate for a single patient acquisition to determine a dose response curve as:

42. The method of claim 41 wherein A is a monotonic function of s, wherein said calibration dose is related to A by a decay factor and excretion fraction, said method further comprising the step of substituting said calibration dose for A.

43. The method of claim 35 further comprising the step of developing a database relating patient characteristics including at least weight, body mass index, sex, disease, diabetes state, and anatomical position to optimal dose and peak SNR.

44. The method of claim 43 further including the step of statistically correlating said optimal dose for each scan completed into said database.

45. The method of claim 43 further including the steps of:

(a) comparing said patient characteristics for a new patient with said database; and

(b) estimating said optimal dose for said new patient.

46. A method for improving clinical data quality in Positron Emission Tomography (PET), said method comprising the steps of:

(a) injecting a patient with a calibration dose of a radiopharmaceutical, said calibration dose being a fractionated portion of a patient dose of the radiopharmaceutical;

(b) pre-scanning the patient;

(c) determining a trues rate (T) by;

$T_{\mathrm{pat}}<br><br>\left(s\right)=\frac{T_{\mathrm{pat}}<br><br>\left(s_{\mathrm{pat}}^{\mathrm{meas}}\right)}{T_{\mathrm{phant}}<br><br>\left(s_{\mathrm{pat}}^{\mathrm{meas}}\right)}<br><br>T_{\mathrm{phant}}<br><br>\left(s\right);<br><br>$

(d) determining a randoms rate (R) by;

$R_{\mathrm{pat}}<br><br>\left(s\right)=\frac{R_{\mathrm{pat}}<br><br>\left(s_{\mathrm{pat}}^{\mathrm{meas}}\right)}{R_{\mathrm{phant}}<br><br>\left(s_{\mathrm{pat}}^{\mathrm{meas}}\right)}<br><br>R_{\mathrm{phant}}<br><br>\left(s\right);<br><br>$

(e) determining a scattered events rate (S) wherein;

${\left(T+S\right)}_{\mathrm{pat}}<br><br>\left(s\right)=\frac{{\left(T+S\right)}_{\mathrm{pat}}<br><br>\left(s_{\mathrm{pat}}^{\mathrm{meas}}\right)}{{\left(T+S\right)}_{\mathrm{phant}}<br><br>\left(s_{\mathrm{pat}}^{\mathrm{meas}}\right)}<br><br>{\left(T+S\right)}_{\mathrm{phant}}<br><br>\left(s\right),$

wherein s is a total measured singles rate;

(f) determining a singles rate

$s_{\mathrm{pat}}^{\mathrm{meas}};<br><br>$

(g) determining an optimal dose of the radiopharmaceutical using T_pat, S_pat, R_pat and

$s_{\mathrm{pat}}^{\mathrm{meas}};<br><br>$

(h) determining a remainder dose as a difference between said optimal dose and said calibration dose; and

(i) injecting said patient with said remainder dose.

47. The method of claim 46 wherein T(s) and R(s) are measured for a given scanner at at least one value of s using a phantom.

48. The method of claim 46 wherein T(s) closely approximates (T+S)(s) in shape and wherein (T+S) defines a net trues rate, wherein (T+S) is measured directly, and wherein said trues rate is estimated by subtracting an estimate of a mean scatter distribution.

49. The method of claim 46 wherein said randoms rate is estimated using one of a group of methods including at least estimating said randoms rate directly from a delayeds sinogram, estimating said randoms rate as a fraction of a total delayeds rate assuming a uniform distribution and using an attenuation mask, and estimating said randoms rate from a spatial distribution of said singles rates.

50. The method of claim 46 further including the step of estimating a signal-to-noise ratio (SNR) as a function of said singles rate.

51. The method of claim 50 further including the step of relating said calibration dose to said singles rate by calibrating a phantom measurement of activity compared to said singles rate for a single patient acquisition to determine a dose response curve as:

52. The method of claim 51 wherein A is a monotonic function of s, wherein said calibration dose is related to A by a decay factor and excretion fraction, said method further comprising the step of substituting said calibration dose for A.

53. The method of claim 46 further comprising the step of developing a database relating patient characteristics including at least weight, body mass index (BMI), sex, disease, diabetes state, and anatomical position to optimal dose and peak SNR.

54. The method of claim 53 further including the step of statistically correlating said optimal dose for each scan completed into said database.

55. The method of claim 53 further including the steps of:

(a) comparing said patient characteristics for a new patient with said database; and

(b) estimating said optimal dose for said new patient.