Updated February 2007

Mal's Musings

Malcolm A Traill

IN  THE  PUBLIC  INTEREST
Updated 18/2/2007

 

NH & MRC REVIEW

 

Chapter 5: Audit of patient medical records (pp 87- 103)

 

“. . . the terms of reference to assess the therapeutic effectiveness of microwave (UHF) cancer therapy.”

(Abbreviations as used in the document will be applied in the following analysis)

 

 

General Comment:

  1. Much of the text involves prolix clerical procedural information of peripheral interest only
  2. Appendices were mislabelled and unclear (eg Appendices 5 & 6, probably being 15 & 16)
  3. The computer program used for the probabilistic record linkage was not specified.
  4. Analysis was also restricted to the comparison of RT with RT+UHF*, since there were clearly too few treated with UHF without RT.” (p. 91) The number for UHF without RT was 53, which some may not consider to be “clearly too few.” No other justification was provided for this decision. *UHF pre ~1990 was usually combined with RT and did not have the GBA.
  5. No “control”/comparison/reference group with no specific therapy was entertained (negative control)
  6. No “control”/comparison/reference group having RT alone was entertained for the “any invasive” group. Thus, the study of “any invasive” cancer was to compare (Tables 40-44) RT+UHF (apples) against UHF+GBA (oranges), with no RT or a negative control group !
  7. Only for the bladder carcinoma diagnosis were there comparison groups of any potential use provided – RT alone (n=34), RT+UHF (n=12) and UHF+GBA (n=18). The non-RT numbers are becoming low.
  8. Emphasis seemed to be placed upon mean estimates, when both mean and median would seem more appropriate.
  9. Table 39 has excluded 9 UHF patients “Excluded because too few numbers.” Whether these were UHF only or UHF+GBA is not clear. A decision by the Committee to exclude these is not open for assessment. The tumour types were not specified.
  10. Details on the statistical checking regarding the proportional hazards model (p 92) were not presented.
  11. Percentages are generally based upon group totals including an “Unknown” class. This tends to dilute the classes with known outcomes, and is conservative. The “unknown” classes have been deleted from assessments.

Results Section

  1. Patients’ demographics (Table 40)

Patients with invasive bladder cancer treated with RT alone were more likely to be newly diagnosed whereas patients treated by RT+UHF or UHF+GBA were more likely to have recurrent disease, making comparisons difficult.” (Emphasis added). Many may consider the word “difficult” in such a statistical context to mean “invalid.” So, the Committee was comparing Pink lady apples with Cider apples and Crab apples.

Constructing a χ2 matrix for bladder cancer (Groups A-C):

            Recurrent                                 No recurrence

            Metastasis [“predicted”]           No metastasis               Total

            RT                   6                                  28                                34

            RT+UHF         9                                  3*                                12

            UHF+GBA      13                                5                                  18                   

RT+UHF      * 22        [5.3]                 8          [24.7]  

UHF+GBA

*Groups combined because the individual values are too small

χ2 Test: P = 1.3 x 10-15 highly significant

Similarly, for “any invasive” cancer (Groups D & E):

                        χ2 Test: P = 0.000241, highly significant

So, without considering other parameters, all the groups are significantly different in important clinical features such as recurrent &/or metastatic tumour.

  1. Patients’ tumour characteristics (Tables 41 & 42)

For the “any invasive” cancer and combining (comb.) the figures for prostate, digestive and melanoma & skin (because some numbers were less than 5):

            RT+UHF (n=56)                      UHF+GBA (n=49)

            Number            Rest                 Number            Rest [“predicted”]    [Total]

Breast  21                    35                    6   [18.4]          43  [30.6]                     [49]     

Lung     6                      50                    6   [5.25]          43  [43.75]                   [49]     

Comb.  12                    44                    17 [10.5]          32  [38.5]                     [49]

χ2 Test: P = 9.04 x 10-5highly significant

·        Bladder carcinoma. Comparisons in text and in the Table 42 do not define the bladder wall +/- “adjacent tissue” in accordance with the definitions provided in Attachment 2 to Appendix 16 (T staging) p.239. The reader is not informed in text that the numbers of patients/n, (%), with “disease located to the bladder or invading into adjacent tissue” was – the staging in the table below is an interpretation Table 41, as continued on p. 96 :

Stage               RT(n=34)        RT+UHF(n=12)          UHF+GBA(n=18)

≤T2b                25  (74%)        8  (67%)                      13  (72%)

T3-4                5    (15%)        3  (25%)                      0    (0%)

T5                    3    (9%)          1  (8%)                                    5    (27.8%)

Unknown         1

The numbers in many of the groups with Stage ≥3 are so small that they are not able to be treated reliably statistically. All one can say, looking at the figures, is that distant metastases (T5) are appreciably more highly represented proportionately in the UHF+GBA group, and would be expected to be more refractory to treatment and have worse parameters for all catecorizations studied.                                                     

·        Any invasive carcinoma. Having noted appreciable differences in tumour types: Group D (RT+UHF) had as highest, 21(38%) of 56 patient with carcinoma of the breast, whereas group E (UHF+GBA) had digestive 8 (16%) of 49. The Committee noted that metastases made “treatment comparisons difficult.” (See earlier.) There are no details of the “other” malignancies, which comprised 17 (30%) of 56 in group D and 20 (41%) of 49 in group E. The Committee makes no comment on why there should be these differences (see later).

  1. Patients’ treatment characteristics (Table 42)

The treatment groups differed making comparisons difficult.” (See earlier.) “. . . patients treated with RT alone were less likely to have had chemotherapy (6%) compared to all other groups (27-56%).” However, the other groups in the table ranged between 44% - 73%, so the origin for their figures is not clear. The median doses for RT (curative intent) in the RT alone group was 60Gy (a rather high dose), whereas, for those having RT+UHF (curative intent) it was 51Gy, range 24 – 55Gy (n=8). With such small patient numbers, one low dose of 24 Gy which, whilst it may have started with curative intent, would have to be considered too low for any cure, and could pull the median dose down (assessment of scatter is not provided). An outlier like that (and any others) should have been explained. Having noted the dosages in Gy, the Table does not mention Gy, despite directing the reader to Table 42 in the sentence dealing with the RT dosage !

·        Bladder carcinoma. After having noted 92-94% “curative intent” for the RT groups, with the UHF+GBA 78%. “Residual macroscopic disease was present in 56%, 33% and 22% . . . ” across the Table. There seems the distinct possibility that the 22% without “curative intent” was also the 22% with “residual macroscopic disease” in the UHF+GBA group, but this is not made clear.

·        Any invasive carcinoma.The treatment intent was curative in [34 of 56] 61% for group D (RT+UHF), but only [9 of 49] 18% for group E (UHF+GBA).”  A χ2 matrix:

Curative          RT+UHF (n=56)          UHF+GBA (n=49)

                        Yes      No                   Yes      No

                        34        22                    9          40

“Predicted”                                           29.75   19.25

            χ2 Test: P = 1.301 x 10-9highly significant 

                        The Committee did not point out that the “Resection (no residual macroscopic)” for the two groups were 21 (38%) and 13 (27%) respectively, only noting that with “Resection (residual macroscopic)” the disease figures of 6 (11%) and 8 (16%) respectively. (The numbers do not add up to the group totals, because they refer to those who had prior surgery.)

Without Residual       RT+UHF (n=27)          UHF+GBA (n=21)

                        Yes      No                   Yes      No

                        21        6                      13        8

“Predicted”                                           16.33   4.67

            χ2 Test: P = 0.081,  not significant, but a trend which hardly gives confidence that there are comparable features.                      

·        Best 10. Since this group is more heterogeneous than the other groups, and lack any form of control positive or negative reference group, they can only be treated as ten isolated case reports, to be assessed on the basis of likelihoods in clinical experience. The committee does not make that point, merely listing that 50% had curative intent, 30% previous RT and 30% prior chemotherapy.

  1. Treatment toxicity and disease symptom control (Table 43)

So what is “toxicity” ? Sadly, the Review does not define this word. In the protocol instructions on p.235, those collecting data are asked to “Please document any treatment-related toxicities, ie signs and symptoms occurring during the study treatment . . .” “The grading of ‘mild,’ ‘moderate,’ and ‘severe’ is based upon the Common Terminology Criteria (http://ctep.cancer.gov/forms/CTCAEv3.pdf) and correspond with grades 1,2,3 and 4 on this scale.” When there is reference to this site, the term “toxicity” (or “toxicities”) does not appear: the internet site refers to “Adverse Event” (“AE”). Recourse to the Macquarie Dictionary (1981) states :

            “Toxicity, toxic quality; poisonousness.”

Toxicity” would then seem to convey the implication that the treatment is harmful, as a poison, which then carries a subjective and pejorative, emotive connotation. “Adverse Event,” on the other hand is more objective and less pejorative. The inclusion of such a term as “toxicity” would seem inappropriate and possibly/probably biased.

Patients treated by RT+UHF appeared to have a higher incidence of moderate or severe toxicity or toxicity requiring hospitalization.” (p. 98)

Not only has the Committee failed to define “toxicities” and give examples but, sadly, the Committee also has not presented the “toxicities” in a detailed tabular form, with specific groups and types of “toxicities” presented for you, the reader to assimilate. The numbers were sufficiently small that such a presentation could have been undertaken easily, and would have been more useful than the mass of irrelevant references collected by the outsourced agency. The “toxicities” for the UHF&GBA group is of particular interest. That is because UHF+GBA is so relatively free of “toxicities” (in the experience gained at Fairfield, Victoria). We have to wonder just what these “toxicities” could be, who found them and what criteria were used. The possible known “toxicities” that might be considered are :

                                                                          Medical

Toxicity              Experience  Frequency        importance    Grade

Venepuncture (GBA)

Pain

Inevitable (less with PICC)

Trivial

Mild?

GBA vein irritation

Pain

Usual

Trivial

Mild?

GBA extravasation

Pain

Occasional

Small

Mild

Taste GBA

Cabbage

Inevitable

Trivial

Mild?

Heat

Sweat

Inevitable

Trivial

Mild?

Amino acid deficiency*

Cerebral symptom

Rare

Minor

Mild

Hypercalcaemia

Ataxia

Rare

Serious

Severe/Hospital

Haemolysis*

Anaemia

Very rare

Serious

Severe/Hospital

*These were reported by Holt. The Thalassaemia conclusion is probably incorrect; more likely paroxysmal nocturnal haemoglobinuria (see Submission to the Inquiry)

The top five “Toxicities” (yellow) could hardly be related to “poisonings,” just as pain after an appendicectomy could hardly be regarded as “toxicity.” Of the rarer “Toxicities” (in orange), only the Haemolysis would seem to be an appreciable toxicity against non-cancerous cells. The amino acid deficiency symptoms were never seen at Fairfield, and seemed to be noted by Holt when using a different GBA. The hypercalcaemia would seem to be one of the most convincing pieces of evidence that the UHF had an effect upon the tumour concerned (large cell lymphoma). The chances of malignant hypercalcaemia developing by chance in the second week of treatment for two patients treated close together would seem to be extremely unlikely, and could be likened to the Tumour Lysis Syndrome; the breakdown products of the tumour being the “toxicity” factor. With only rare “toxicities” ever observed at Fairfield in the UHF+GBA group, why then, should skin warming for three times 6 – 8 minutes (in quick succession) add “toxicities” to the patients receiving RT+UHF ? Your guess will be as good as any other’s. There is a lack of scientific and biological plausibility regarding the “toxicities.” 

So, dear reader, when the Committee lists “toxicity” with UHF+GBA as:          Number of “toxicities” 0            17 (35%)

1                    8 (16%)

2                    13 (27%)

3                    6 (12%)

≥4        5 (10%)

                        Maximum degree of “toxicity”  0          17 (35%)

(n=49)                         Mild      9  (18%)

                                                            Mod.    22 (45%)

                                                            Severe  1  (2),

just what is meant ? What were they ? What were the thresholds ? Who documented them in the clinical files ? Who interpreted them from the clinical files ? Were there other “toxicities” which were never seen at Fairfield that they have discovered ? These are secrets that only the Committee knew, and the Committee has not been inclined to tell you or anybody !

The Committee seemed impressed with the “toxicities” found in the combination of RT+UHF and concluded on pp104-5 “. . . patients treated with UHF+RT had the highest toxicity (75% moderate or severe), possibly consistent with a radiosensitizing effect from the UHF).”  (The Committee seems prepared to concede that the UHF might have a radiosensitizing effect!) Surely knowing just what these “toxicities” were would have been of value in assessing the Review’s findings.

                        Just for fun, the χ2 tests were applied to the figures. However, considering that not only was the Committee comparing Oranges with Apples, but the latter were subdivided into Pink Lady Apples, Cider Apples and Crab Apples (and the Bananas were not compared with anything). There is the old saying from the computer world of the 1970s “rubbish in, rubbish out:”                                                 

Toxicities        _______Bladder______                     Any invasive

Number            RT(n=34)         RT+UHF(n=12)           RT+UHF(n=56)

            0          3 (9%)             0                                  7 (12%)

            1          9 (26%)           2 (17%)                       15 (27%)

            2          10 (29%)         3 (25%)                       9 (16%)

            3          6 (18%)           2 (17%)                       7 (12%)

         ≥ 4          6 (18%)           5 (42%)                       18 (32%)

Max. degree

            Nil        3 (9%)             0                                  7 (12%)

            Mild     16 (47%)         3 (25%)                       9 (16%)

            Mod.    10 (29%)         6 (50%)                       35 (62%)

            Severe  4  (12%)          3 (25%)                       1  (2%)

            Hosp.   1  (3%)            0                                  1  (2%)

            Stop     0                      0                                  3  (5%)           

Note the absence of the RT alone in the “Any invasive” class and the small size of the “Bladder” RT+UHF. To construct a χ2 matrix, the two RT+UHF figures will be combined :

Toxicities        Bladder                                   Any invasive

Number            RT(n=34)         “Predicted”       RT+UHF

            0          3                      6                      7                                              1          9                      18                        17

            2          10                    20                    12

            3          6                      12                    9

         ≥ 4          6                      12                    23

                                    Total    68                    68

            χ2 Test: P = 0.00652, significant

Max. degree     (n=33)              “Predicted”       RT+UHF(n=67)

            Nil        3                      6.1                   7

            Mild     16                    32.5                 12

            Mod.    10                    20.3                 41

            Severe  4                      8.1                   7 

            Figures too small                                  

Total    67                    67       

            χ2 Test: P = 1.695E-07,  very significant

As presented, and using the data which cannot be validated or explained, there is indicated a significant “toxicity” difference between the RT alone and the RT+UHF. (Why the treatment was stopped has not been explained  and that group is disregarded here.) Notable, is the slight increased severe/hospitalized “toxicity” for the RT alone, with the major differences appearing between the mild and moderate groups, thereby differing from the conclusion drawn by the Committee, that the moderate or severe/hospitalized groups had a higher incidence.  (Creative bias using flawed data.)        

RT+UHF (n=32)         “Predicted”       UHF+GBA (n=25)

                                    Yes      25                                19.5                 14

                                    No       7                                  5.5                   11

                  Total    32                                25                    25

χ2 Test: P = 0.0079,  very significant

This result is expected, because the combination of RT+UHF would be regarded as likely to be more effective than the UHF+GBA. (Biological plausibility.) However, the place for the UHF+GBA is frequently for patients who have failed, or unable to have radiotherapy (for all sorts of reasons; the comparison being not clinically relevant), so that the real issue is not a comparison between the two types of treatment, but a) Does the UHF+GBA seem to have an effect (possibly) then, b) Does it assist patients and, with a desirable result in 14 of 25 patients (56%), the symptomatic relief validity seems possible. (Clinical plausibility.) This assessment of symptom relief is likely to be more valid than the nebulous, specious and suspect “toxicity” comparisons earlier.

e)      Treatment Outcome (Table 44) The Committee notes the difficulties encountered in assessing follow-ups, noting “However, for many patients an assessment of response could be made towards the end of therapy or at their first follow-up visit or from a letter from their referring oncologist.” This, in the light of the likely effects of the UHF+GBA on tumours, as were provided in the Submission to the Review, is simplistic, in that the Committee has not taken into consideration that the effects may be merely to slow the intrinsic growth rate of the stem cells the tumours. Such changes may not become apparent for months, and can now be likened to the documented effects of Interferon-α on Chronic Myeloid Leukaemia, where the “leukaemic/tumour stem cells” (0.1-1% of the abnormal cells) are principally affected, but with little immediate change to the bulk of the more differentiated leukaemic/tumour cells (see elsewhere on the website).

·        Bladder carcinoma. The lack of the control negative group is a major problem, added to which are the differences in macroscopic condition of the tumours. In addition, the “Not applicable/not known” category for the “Tumour response within three months of treatment” with UHF+GBA  listed Complete responses, 2 (of 12 recorded responses, 16.7%), Partial response 1 (8%) and Stable disease 1 (8%): the “desirable” total = 4 (33%). The “desirable” comparisons between the RT and RT+UHF are 21 (95.5%) of 22, and 6 (67%) of 9 respectively. Recall that the RT and RT+UHF groups had significant differences with the RT+UHF group having the more severe/advanced forms of the disease :

 

                                                      RT (%)   [n]                RT+UHF (%) [n]

Invasion of adjacent tissue. . .”   5   (15%) [33]             3    (25%)        [12]

 “macroscopic”                              23 (68%) [34]              11  (92%)        [12]

Without comparable groups and a control negative group, the figures are effectively meaningless.

·        Any invasive cancer. The figures for this group have been examined already as a follow-on discussion to the Submission to the Review, which indicated that, if UHF+GBA were compared to a control negative “guestimate” group as provided by Professor Fox, the differences were highly significant. Notable, was the incidence of “stable disease” 24 (61.5%) of 39. Comparison between the RT and RT+GBA groups is not clinically relevant, because the former would normally be done, if feasible: the role for the UHF+GBA is largely for those there RT is not appropriate or possible. The information available (p. 95) was that the RT group had 38% carcinoma of the breast (usually with some responsiveness to ionizing radiation), whereas within the UHF+GBA group, 16%, the largest subgroup, came from the “digestive tract” (not usually very responsive to ionizing radiation, unless squamous cell carcinoma was included and, we can assume it was not) which would seem to indicate that there was a very mixed collection of malignancies. Sadly, the failure to provide the proportions of malignant types &/or recognized responsiveness to ionizing radiation, leaves a large gap for comprehension or interpretation of the results.

·        Best 10. (Tables 44 & 45) The concluding sentence (p.100) “Nine (of 10) patients had complete remission or stable disease at last follow up.”  One could hardly object to such clinical plausibility, but the inability to compare the figure with any reference group limits the study’s usefulness. Why was the term “salvage mastectomy” used for a patient said to have DCIS (yet under the column “Degree of spread” is written “localized”) ? (Just because DCIS was found and no invasion identified in the sections examined[1], does not mean that there was no invasion somewhere, possibly before the UHF+GBA.) The “Best ten” (p. 101) is basically a repetition of the earlier section. There was no attempt to comment upon the patients and their figures.

·        Western Australia Cancer Registry Analysis (p. 101, Table 46). Here, the Committee compares the survivals at 5 and 10 years obtained for the major disease groups on the Registry, and matches that with the survivals at these times for those from Holt’s treatment groups. Again, this comparison does not address the comparison between patients not receiving any form of active therapy, being the comparison that is clinically relevant. The numbers extracted for the cancer sites considered, as “predicted” compared with those from Holt’s practice, with the mean age differences [RT+UHF] – RT (Δ age) :

 

 

Site                  RT (WACR)   “Predicted”    RT+UHF (Holt)          Δ age

            Bladder            95                    19.6                 13                                +5.1

            Breast (F)         387                  79.9                 49                                +2.5

            Colorectal        113                  23.3                 54                                +0.5

            Head & Neck  119                  24.5                 22                                -2.8

            Lung                 325                  67.1                 103                              +1.6

            Lymphoma       150                  31.0                 15                                +6.7

            Prostate            221                  45.6                 35                                +2.7

                        Total               1410                291                  291             Median =  +2.5

χ2 Test: P = 3.62E–16,  very highly significant

Not only would there seem to be one or more unknown &/or unspecified selection factors operating that favour in particular colorectal and lung cancers for Holt’s series, but the mean ages are different, with the median age difference +2.5 (-2.8 to +6.7) year for Holt’s series. The reasons that may underlie these differences are not mentioned nor discussed. Not only are we comparing oranges with apples, but the items are of different ages and proportions.

            Table 47 deals with the number who had died, or survived 5 and 10 years. Given that the groups can hardly be called comparable, for reasons which have not been exteriorized, this, and Table 48 & 49, which attempts to legitimize the comparisons by expressing the results in terms of “Hazard ratios” and “Adjusted survivals”, must be of very questionable value. (The former expresses a measure that compares the “mortality rates” [p. 93]. Just what parameters were used to derive the “Mortality rates” are not stated or exteriorized clearly; presumably death, either within a certain time frame or by survival post diagnosis or treatment (which ?). Whilst standard deviations are implied in the results, they are never exteriorized, as for ranges, [except those under Table 40, p. 94]. Interestingly, despite the selection and non-comparable groups, the statistics for bladder, colorectal and head & neck cancers are presented as showing no significant difference. The “Adjusted survivals,” based upon the previous questionable database, presents a possible advantage for RT+UHF at 5 year for Bladder and Head & Neck cancers, a difference that persists to 10 year.

  1. Discussion

                     i.            The Committee’s Discussion (p. 104 - )

o       The two-part form of the study is outlined, comparing case notes of 179 patients and “. . . a detailed matching study with the Western Australian Central Cancer Registry, which compared the survival outcome of 1701 patients with seven different cancer sites with RT or RT+UHF.” The number of patients from the Registry listed in Table 46 is 1410, providing only the mean age at diagnosis. Sadly, the criteria by which there is selection are not detailed, not is there a presentation of the comparative features of the groups under examination, as seen in most publications matching groups. There is no indication of the comparative scatter in each group – the mean, standard deviation and median should have been presented for all groups in the comparison.

o       The Committee notes Holt’s stated belief that bladder cancer is “. . particularly sensitive to treatment with RT+UHF” and quotes from his earlier work from 1988 in which he treated patients with Stage T1-T2 disease. However, in Table 44, only 24% of the RT group and 67% of the RT+UHF group had Stage 1-2 disease, so the comparisons and comments of the third paragraph may be misleading.

o       The long term toxicities were not well recorded . . .” These were never specified. They go on to list and compare the undefined and unspecifiedtoxicities,” with reference to those seen with “invasive bladder carcinoma,” which has not been specified in the Tables (“Localized” is taken to mean “non-invasive” cancer here). They state that “UHF+GBA had the lowest toxicity - 37% of patients with invasive bladder carcinoma . . . had moderate to severe toxicity.” (Emphasis added)  Table 41 lists 13 (72%) as having “Localized cancer.” Similar conclusions are drawn for the RT and RT+UHF groups, with the latter scored 9 (75%) moderate to sever toxicities, apparently assessed on a weekly basis (were the patients in the RT arm similarly assessed ?). The Committee comments “. . . possibly consistent with a radiosensitizing* effect from the UHF.” * “Radiosensitizing” refers to increasing the sensitivity to ionizing radiation; eg X-ray. “Despite the higher toxicity, patients with invasive bladder cancer treated by RT+UHF had lower disease symptom control rates than patients treated with RT.” Not only was the mean age of the RT+UHF group some 5.1 year greater than the RT group, but they had a greater proportion of invasive disease cases 4 (33%) of 12 compared to 8 (24%) of 33, (the Tables do not tell us how advanced the diseases were in each category) and the 7 (64%) of 11 had symptoms at the outset compared to 6 (15%) of 33 in the RT group. Such comparisons are not comparable because of the high incidence of symptoms in the RT+UHF group at the start – they had different attributes and were sicker, with more advanced disease. Much the same can be concluded for the “Any invasive” RT+UHF and UHF+GBA groups.

o       The ten best cases. For us, the readers of the Review, to be able to assess these cases, we need to have the detailed information, so that they can be assessed on a case-by-case basis. This requires enough detailed points to form a longitudinal, detailed study on each individual. No-one can form any real opinion (which is what that is involved) from the potted few words provided, as in Tables 41 and 42. One would like to know, for example, how the patient with DCIS presented ? Why did she have the UHF+GBA (the GBA is presumed) ? How thorough was the search for invasion ? Why was there a “salvage mastectomy ?” What is the meaning of “salvage” here ? What did the histology of the mastectomy specimen show ? What did the post operative imaging on the “atypical malignant meningioma” show ? What were its characteristics (eg grade of malignancy, type of invasion etc.) ? What were the initial histology, imaging and diagnostic features of the mesothelioma ? Spontaneous remissions are generally recognized; some malignancies such a malignant melanoma and neuroblastoma have particular reputations for being able to do this – but it is still very rare. To assess the probabilities of these and the tumour behaviour (for clinical plausibility) requires more clinical information over the treatment interval and subsequently – it cannot be done on what is provided, which presents the patients as frozen in time. Readers are left in the dark. The impression gained when at Fairfield, both from verbal accounts from Western Australia, and in Holt’s writings, was that Astrocytoma malignancies formed a “gratifying” group to treat (particularly regarding the disappointing responses to conventional treatments). However, none is represented in this “ten best” group.

o       One of the strengths of the study was the meticulous audit process undertaken by experienced data management and clinical staff with expertise in clinical trial design.” Dear reader, should it be for the Committee to tell you just how marvellous they were with such self-adulation ? Respect for the work done should come from the results of the study, with meticulous detail given to information presentation, such as clinical summaries, clinical classifications and comparisons, and adequate information, such as the mean, standard deviations and median for all groups listed. Economy of paper use did not seem to be an issue.  In the current analysis of the Review there have been listed what would seem to be apparent omissions in these respects. Many may feel that the open expression of such fulsome self-praise arouses suspicions – that there are being covered-over inadequacies and features which should be open to see.

o       The Committee comments upon the limitations of the study as they see it. In particular, the difficulties in applying standardized criteria (eg the CR, PR, SD and PD classifications.) They then remind us all that “. . the UHF treatment was and still is investigational in nature . . . . .” The use of the word “investigational” seems strange and pejorative. The term “under assessment” would seem more appropriate.

o       There is mention of the problems in classifying patients according to the CR, PR, SD and PD classifications, particularly with regard to collecting “objective” data. “In such cases, the determination was made following discussion between the data managers (both experienced oncology nurses) and data verifiers (both experienced oncologists).” The subjective interventions are not presented in the data tables so, dear reader, we have to take their word for everything.

o       There is then a discussion on the use of a comparison group seen and treated at the Perth Radiation Oncology Centre. Having announced that “The analysis showed a survival disadvantage for patients treated with RT+UHF for four of the seven cancer sites (breast, lung, lymphoma and prostate) and no significant difference . . .[for the rest].” The Committee then notes that “the groups were not strictly comparable. . . .”

o       The Committee speculates briefly on the reasons for the survival disadvantages (listed above), but notes that the median radiation dosage for the RT alone was 60 Gy in 32 fractions and for the RT+UHF group 51 Gy in 34 fractions (not in the Tables; no mean and standard deviation provided).

o       The Committee concludes “In summary, a meticulous audit of historical patient records did not find any advantage for the addition of UHF with RT in terms of tumour response and symptomatic control. . .  Further, in four sub-types of patients (breast, lung lymphoma and prostate) accounting for over 1200 patients, survival was significantly inferior for patients treated by RT+UHF compared to RT.” There is no mention of UHF+GBA, and the Committee failed to mention the sub-types of bladder cancer and head and neck cancer, which did seem to benefit. (These would have been transitional cell &/or squamous cells carcinomas. None of these sub-types was typically adenocarcinoma.)  

                   ii.            Comments upon the Committee’s study

·        At the heart of the problem, is the failure of the Committee to try to understand how the UHF +/- RT/GBA may act upon tumours. One would imagine that there would be some point in the Committee reading the Submission to the Review from one in Australia, other than Holt, who had applied the treatment (as opposed to the armchair commentators). That was not done, so the Committee failed to read how UHF might act, how it may affect tumour growth and how it seemed to affect tumours, at least with respect to tumour markers. The conclusion reached in the Submission to the Review was that the UHF+GBA seemed to slow tumour growth. With this in mind, the Committee could have planned a study not bound to the traditional CR, PR, SD and PD concepts. This would have allowed the gathering of data from control negative patients – (those who have untreatable cancers/sarcomata) in ways that are acceptable morally (eg by studying the intrinsic growth rate of secondaries in the liver, lungs and nodes of patients for whom chemotherapy +/or radiotherapy are either not applicable or refused). The CR, PR, SD and PD classification was too crude, and involved subjective assessments for classification. Measurements of superficial tumour lumps and by imaging of deeper lumps is easily done, more objective and, to judge by Holt’s writings, was done by him. All results should have expressed tumour volume and its change.

·        The data presentation left a lot to be desired. The presumption seemed to be that the Committee and its staff could do all the assessing of the data and present only the minimum for you, dear reader, to read (but not consider). Not only is there no possibility to study and make one’s own conclusions from the data, because there is too little patient detail presented, but there is a lack of detailed comparisons between the groups, with notable age differences, dose differences, differing percentages of digestive tract disease (notoriously relatively refractory to radiotherapy), differing symptom levels, and (by inference) different sickness levels etc. all expressed without any real indication of the scatter within each group (the standard deviation and median). Yet the Committee considered these aspects exemplary !

·        “. . . the terms of reference to assess the therapeutic effectiveness of microwave (UHF) cancer therapy.” Holt has made two main claims :

1)      That UHF has a beneficial therapeutic effect upon human cancers, augmented by the administration before the UHF exposure, of oxidized glutathione + S-methyl-L-Cysteine Sulphoxide which, for the purposes of this presentation, has been abbreviated by Holt as “GBA.” Holt has used this combination as a monotherapy from ~1990 and during the course of the Review. This was the relevant, extant treatment to study. The Committee’s study design and conclusions could not address whether this treatment provided any benefit or not compared with no treatment at all. The Committee failed under its terms of reference in this respect.

2)      That UHF, when administered before and within 25 minute of the radiotherapy fraction on the day, “sensitizes” the tumour to the radiotherapy. As a result, Holt typically administered an appreciably lower dose of Radiotherapy (eg the median of 51 Gy to 60 Gy, with no indication of the dosage spread, see above). In the study design, the assessment of this claimed “sensitizing” effect was the only aspect of Holt’s major claims that this study could address. Interesting as it may be, sadly, the groups are so poorly matched that no firm conclusions can be made, but lets continue the charade and examine the figures to learn by the mistakes :

a)      Survival. According to Table 49, a survival benefit of RT+UHF is shown for bladder cancer and head and neck cancer. Neither bladder cancer nor head and neck cancer are typically adenocarcinomas, such as breast, colorectal, lung and prostate. Holt and Nelson (1996)[2] pointed out, when commenting on the study by Trotter et al. (1996)[3] that the RT-UHF augmentation was only expected to be apparent when dealing with cancers that are (typically) reasonably sensitive to ionic radiation. Adenocarcinomas are not typically sensitive – they are typically relatively resistant to ionizing radiation. The application of UHF may also be limited in affecting well- to moderately-differentiated adenocarcinomas, because the conduction path through the tumours will be expected to have numerous discontinuities, due to gland and cleft formation, stromal septa and mucus pooling etc. There are, therefore, explanations as to why the UHF may not augment the reduced dose of radiotherapy given for these patients. Conclusions or comments based upon some 3 lymphoma patients on little more information provided, can hardly be valid, should be excluded, and consideration of this subtype will not be pursued further here. Adenocarcinomas which are poorly- to undifferentiated will be expected to have a better electrical conduction path, with the expected cancer cell to cancer cell conductions’ “guestimate”:

 breast > (lung/prostate) > colorectal.

Breast carcinomas would be more likely to fall into the more poorly differentiated class, and the claimed differences between the RT and RT+UHF are less than for the other adenocarcinomas at 5 year,

RT (CI*)                      RT+UHF (CI)

      (Table 49) –    “Breast            72%  (68-76)    versus 56%  (46-69).”

Colorectal      17%  (11-25)    versus 9%   (4-19)

Lung              4.4% (2.7-7.2)  versus 1.5% (0.6-3.9)

Prostate         56%  (50-62)    versus 34%  (23-51)

*the 95% confidence interval (CI)

There is an explanation for the likely reduced benefits of both RT and UHF on adenocarcinomas, the reasons seemingly being scientifically plausible.

Looking now at the bladder carcinomas and head and neck carcinomas, there is stated to be a benefit in coming RT and UHF, which is found despite the lower median dose of the RT in the RT+UHF combination, with the five year survivals :

 

                                                            RT (CI)                        RT+UHF (CI)

(Table 49) –    “Bladder                      20% (13-30)    versus 28% (12-66).”

Head & Neck             36% (28-46)    versus  42% (27-67)

 A median dose of 51 Gy (with no scatter measure provided) would normally be considered a low dose for transitional cell and squamous cell carcinomas (and probably ineffective for adenocarcinomas). Yet despite this, the combination of RT51Gy+UHF performed better than RT60Gy. The conclusion, then, is reasonably clear: based upon the flawed data sets and the Committee’s calculation, the UHF adds the equivalent of ~9+ Gy to the effectiveness of ~51 Gy: Therefore UHF increases the sensitivity of the tumours to ionizing radiation. All this ignores the flaws and faults noted in the patient matching and the documentation that was presented earlier. If we deduce, by examining the Tables, that the RT+UHF patients were sicker, then the benefits added by the UHF would be greater.                  

b)      Symptom control. This topic has been dealt-with earlier. The bladder cancer RT and RT+UHF groups pre-treatment were clearly dissimilar, and there is no point writing more on the topic. 

c)      Toxicity. The information, definitions, classifications and documentation relating to toxicities are so poor, that this morbidity needs to be dropped out. In addition, the Committee’s documentation and conclusions lack biological plausibility. There is no point writing more on the topic.

Conclusion to this part: Looking at the figures provided by the Committee, but mindful of the flawed and specious nature of the presentation, and the biased conclusions, the following points can be proffered as indicators for better and further studies :

1.      Holt’s claim that the UHF, when administered within 25 minute before ionizing radiation appears to increase the sensitivity for those tumour types which are known to be inherently and reasonably sensitive to the ionizing radiation.

2.      The combination of RT+UHF may have a small sensitizing effect in adenocarcinomas that are poorly differentiated or undifferentiated. This may not be enough to offset the reduction in ionizing radiation dosage that Holt has typically applied. A higher radiotherapy dose and more powerful UHF dose for shorter times may be tried.

3.      The addition of UHF to ionizing radiation for well- to moderately-differentiated adenocarcinomas, may not benefit the outcome appreciably for reasons that may be explained scientifically.

4.      By failing to analyse the data appropriately (as has been done here), the Committee has failed in its duty to address, in a professional way, the terms of reference from the Minister for Health on this (the RT/RT+UHF) issue and the UHF+GBA treatments issue.

The Committee’s job is not finished.

5.      The study provides clinical plausibility for treatment benefits to accrue from the combination of RT+UHF in appropriate cases.

 

UHF 434 Mz and Ionizing Radiation:  Scientific rationale.

a)      The probable combined effects upon the centriole(s) were mentioned in the Submission to the Review (see elsewhere on the site). There were quoted references in which the possibility was entertained that an early effect of ionizing radiation may affect the centriole(s), with secondary/later effects upon the nucleus. If this be the case, the assault of UHF upon the centriole(s) may be to augment the damage, although there could be a problem explaining why there should be augmentation when the UHF is administered before (but not after) the ionizing radiation. These effects need not be exclusive :

b)      More recently, there has been attention given to the reasons for resistance to ionizing radiation by tumour stem cells. There are now numerous reviews and articles on the topic, and some predate the conclusion of the Committee’s Review, meaning that well read Committee members should have been acquainted with the work being done prior to the release of the Review in September 2005. This work should have influenced how the Committee planned and carried out its Review but, sadly, seemed not to have.

c)      The major points to come out of the recent work are ;

                                 i.            Tumour stem cells comprise some 0.1-1% of the cell bulk of tumours[4],[5],[6],[7],[8],[9],[10] (studied to date)

                               ii.            They typically proliferate rapidly, possibly provide paracrine stimulation to surrounding cells[11] and stock the tumour with more differentiated cells which then grow more slowly, but make up the bulk of the tumour

                              iii.            Most standard treatments may be impressive in producing tumour bulk size reductions by eliminating the more differentiated cells, but generally they fail to have much impact upon the tumour stem cells which remain to divide and ultimately cause the patient to die; there being little significant benefit to survival by the therapy, despite the therapeutic “debulking.”

                             iv.            One of the therapies which seem effective against leukaemic (tumour) stem cells[12],[13] is Interferon-α. Since there were indications that the UHF-GBA slowed the intrinsic tumour growth rate (see the Submission to the Review), the effect of this UHF treatment may be on tumour stem cells also (UHF not being dependant upon plasma membrane controls), accounting for less impressive tumour bulk reduction but slower growth, and less paracrine stimulation. (This may be illustrated by the Photomicrographs of the breast cancer to found in another part of the website.) Lithium, by initiating a cytokine cascade, (under the right protocols), may affect tumour stem cells similarly.

                               v.            Tumour stem cells seem to be (relatively) resistant to ionizing radiation and most chemotherapeutic agents.

                             vi.            Current thinking is that the ionizing radiation damages DNA, and there are cell mechanisms to detect this damage; initiating signals that cause a halt to further cell division (they activate one or more checkpoints[14],[15],[16],[17], typically that at G2M, but there are others at G1 and S) and give the repair processes time to repair the DNA damage before the steps leading to mitosis can restart and proceed inexorably to cell division

                            vii.            Cyclin B1/Cdk1, a protein kinase, forms at the centrosome[18],[19],[20], and is fundamental to the further cell cycle progression. In order to form this complex, Cdk1 must be as it was when produced by transcription-translation, un-phosphorylated and active. It is normally kept inactive and in readiness by phosphorylation at Threonine 14 and/or Tyrosine by the protein kinases[21],[22],[23] Wee1, Mik1 and Myt1. Wee1 is also known to associate with the centrosome, phosphorylating tubulin and being physically involved in the phosphorylation of the γ-Tubulin Ring Complex, the caps on the proximal (centrosomal) end of the microtubules, which are typically found in the pericentriolar matrix or attached to the subdistal appendages of the mature centrioles.

                          viii.            In order for the cells to divide, the Cyclin B1/Cdk1 needs to be activated, and to achieve this, the phosphatases Cdc25 B/A/C[24] are activated and strip the phosphate groups from the Threonine 14 and Tyrosine 15 of Cdk1, making the Cdk1 active: and then the cell cycle can proceed into mitosis (M).

                             ix.            DNA damage induced by ionizing radiation (radiotherapy) triggers a cascade of events which ends up with the phosphatases Cdc25 B/A/C being inhibited, meaning that the phosphate groups stay on the Cdk1, keep it inactive and block the function of Cyclin B1-Cdk1 – the checkpoint is activated and progress to cell division is halted.

                               x.            UHF may, by means of its effects on the centriole(s) (eg athermal heat, other – see Submission to the Review) affect the balance between the phosphorylation of Cdk1 and its dephosphorylation by the Cdc25s, such that, if Wee1/Mik1/Myt1 are damaged/inhibited, the checkpoint activation steps involving the inhibition of the Cdc25 B/A/C will have little or no effect on the non-phosphorylated Cdk1, which will maintain its activity and the cell will proceed to mitosis without a delay.

                             xi.            To be clinically relevant, the damage or inhibition of the Wee1/Mik1/Myt1 kinases must last during the time of the ionizing radiation and for about another 10 minute pause, so that the alert/repair cascade to activate the checkpoint abates and the cells can proceed through the inactivated checkpoint into mitosis, not allowing time for the ionizing radiation-induced DNA damage to be repaired. This would take approximately 55 minute (based upon Holt’s claim that the UHF must be administered before and within 25 minute of the ionizing radiation).

                            xii.            Cells so affected would proceed into mitosis with the ionizing radiation-induced DNA damage, be held up in mitosis (because of the chromosomal disorganization) and, if they can complete cell division, pass on very defective chromosomes to the daughter cells; damage which may be expected to induce apoptosis or cell death.

                          xiii.            This explanation is non-exclusive also. For example, the role of an earlier member of the cascade for DNA damage repair, Chk1, is found at the centrosome[25] in interphase. It seems to protect Cdk1 from premature activation by cytoplasmic Cdc25B. The latter needs to be activated at the centrosome by phosphorylation of Serine 353 by Aurora A[26]. All these activities may be perturbed by heat (etc.) at the centrosome.

Summary:       There are cellular mechanisms now recognized to be in place, which could be perturbed in ways that may block the G2 checkpoint response to DNA damage. The above explanations attempts to, and possibly provide scientific plausibility to Holt’s claim that the UHF 434 MHz, when administered before and within 25 minute of ionizing radiation (radiotherapy) sensitizes the cancer cells to the latter. There may be other explanations. There is plausibility, clinical and scientific, consistent with, and supporting Holt’s key claims.

 

The Committee failed in its duty to the Minister and the public.

 

 

Copyright © MA Traill, February 2007-02-18

 

Press “Back” to Return

 


 

[1] Azzopardi JG. “Problems in Breast Pathology” 1979, pp 268-271. WB Saunders Company Ltd.

[2] Holt JA Nelson AJ.  Australasian Radiol. 1996; 41(3):317-8

[3] Trotter JM Edis JA et al. Australasian Radiol. 1996 40:298-305

[4] Jones RJ Matsui WH et al. J. Natl. Cancer Inst. 2004; 96(8):583-5

[5] Huff  CA Matsui et al. European J. Cancer. 2006; 42:1293-7

[6] Singh SK Clarke ID et al., Cancer Research 2003; 63:5821-8

[7]Al Hajj M Wicha MS et al. Proc. Natl. Adad. Sci. (USA) 2003; 100(7):3983-8

[8] Ponti D Zaffaroni N et al. European J. Cancer 2006; 42:1219-24

[9] Hemmati HD Nakano I et al. Proc. Natl. Adad. Sci. (USA) 2003; 100(25):15178-83

[10] Zhou B-B S Elledge SJ et al. Nature 2000; 408:433-438

[11] Assanah M Lochhead R et al. 2006; 26(25):6781-90

[12] Passegué E Jamieson CH et al. Proc. Natl. Adad. Sci. (USA) 2003; 100(Sup 1):11842-49

[13] Hasford J Pfirmann M et al. Haematologica 2005; 335-40                                                

[14] Xu X Tsvetkov LM et al. Molec. Cell Biol. 2002; 22(12):4419-32

[15] Becker-Jensen S Lukas C et al. J Clin. Biol. 2006; 173(2):195-206

[16] Myers JS Cortez D. J. Biol. Chem. 2006; 281(14):9346-50

[17] Boddy MN Furnari B et al. Science 1998; 280:909912

[18] Lindqvist A Källström H et al. J. Cell Science 2004; 417(21):4979-90

[19] Lindqvist A Källström H et al. J. Cell Biol. 2005; 171(1):35-45

[20] Jackman M Lindon C et al., Nature Cell Biol., 2003; 5(2):143-8

[21] Stumpff J Kellogg DR et al. Current Biol. 2005; 15:1525-34

[22] Booher RN Holman PS et al. J. Biol. Chem. 1997; 272(25):2230-6

[23] Jin Z Homola EM et al. Development 2005; 132(18):4075-85

[24] Donzelli M Draetta G. EMBO Reports 2003; 4(7):671-7

[25] Kramer A Mailand N et al. Nature Cell Biology 2004; 6(9): 884-91

[26] Dutertre S Cazales M et al. J. Cell Science 2004; 117:2523-31