Figure 3. Forest Plots on Bevacizumab efficacy(36)
Pooled data (a) PFS (b) OS from trials on use of Bevacizumab as both first and second line treatment. The diamond at the last row of each group shows overall effect and those crossing the middle line denote insignificant results.
Lombardo.G et al collated and analysed the major studies for both first and second line treatment and results are presented in Figure 3(36). For both settings, pooled results favoured BEV treatment for PFS(36). A noteworthy amelioration in PFS was observed, with a hazard ratio of 0.7 and 0.52 for new and recurrent respectively(36). However, the trials did not corroborate that BEV confers a survival benefit from the summarized OS(36).
Some studies examined the reason for the unsuccessful attempt of BEV in improving OS(37). One hypothesis is that tumour cells adapt by diverting to a separate pathway for angiogenesis, so the effect of drug is only transient(38). Mice models revealed that inhibition of VEGF to achieve anti-angiogenic effects causes vessel cooption, thus metastasis(39).
Furthermore, present imaging techniques cannot accurately determine extent of tumour cell death(40). Contrast-enhancement MRI could give a higher estimate than actual, using new criteria e.g. RANO are means to enhance measurements(40). PET scan is also proposed but it not a direct measure of cancerous tissues loss(40).
A. Anti-?-amyloid immunotherapy
The amyloid hypothesis proposed that fragments of ?-amyloid released from the cleavage of amyloid precursor protein by ?- and ?-secretase join to generate soluble oligomers(41). These oligomers cluster to form fibrils and eventually plaques causing dysfunction of synapse and neuronal death(41). To target these neurotoxic proteins, a range of ?-amyloid antibodies has been developed but none of these drugs is currently indicated for treating AD(42). The efficacies of a shelved drug, Bapineuzumab (BAP) and a more encouraging candidate, Aducanumab (ADU) are covered in the following section.
Mechanism of Action:
BAP and ADU target different structural stages of ?-amyloid(43). BAP binds to the protofibrils, specifically the N-terminus(43), while ADU binds to the oligomers and protofibrils(44). However, the mechanism of neurotoxic ?-amyloid elimination remains ambiguous(45). Proposed theories consist of the phagocytic activity of microglia, neutralizing effects of antibodies and reduction of ?-amyloid in peripheral sites of the body to encourage clearance from brain(45). However, the peripheral sink theory has been disproved in some studies(46).
Janssen and Pfizer funded every trials reported below. All the trials were double blinded and placebo controlled, and explored patients with mild to moderate stages of AD, while trials apart from the Salloway et al Phase III studies had small cohorts.
Salloway et al (2009) Phase II
The trial examined effects of incremental doses of BAP compared to placebo(47). The primary end-points for cognitive (ADAS-Cog) and functional states (DAD) were both not met(47). 0.15mg/kg BAP was noted to cause the greatest change but only under the ADAS-Cog scoring(47). The study was originally intended for safety assessment but revised based on findings from a prior phase I trial(47). Details of the rationale for study alteration were not documented. Study was underpowered for evaluating efficacy and attrition rates were 26% in BAP arm and 21% in placebo arm(47).
Rinne et al (2010) Phase II
Instead of using cognitive tests, the change in ratio of carbon-11-labelled Pittsburgh compound B (11C-PiB) retention in cerebrum to cerebellum from baseline was chosen as the primary outcome(48). The biomarker 11C-PiB reflected the deposition of ?-amyloid in the brain and a decrease in this marker, that was significant compared to placebo, was exhibited in all the scanned regions of the brain(48). However, the change in recorded ADAS-Cog values was not significant(48), suggesting that benefits on cognitive performance was minimal. The 0.15mg/kg dose identified to be most effective formerly was not used in the trial(47).
Salloway et al (2014) Phase IIIa and b
Two international studies on patients with opposing APOE ?4 allele status were conducted and the same scales as that in 2009 trial were adopted as primary outcomes(49). APOE ?4 was reported to be associated with greater AD risk and accumulation of ?-amyloid(50), so in theory, anti ?-amyloid immunotherapy should impact this subgroup more. Tau levels in cerebrospinal fluid and rate of ?-amyloid storage declined significantly in carriers as expected(49). Nonetheless, neither groups received significant clinical benefits from the treatment(49). The DAD score for patients without APOE ?4 only supported the use of BAP if the cut off score for determining the mild AD is greater than 20 instead of 21(49).
Summarising data from the four trials (Figure 4), BAP was considered as an unfavourable drug candidate for treatment of AD(51). It did not cause significant improvements in ADAS-Cog(51). As I2 is 54% for the DAD data, the studies were heterogonous, so the insignificant effect on DAD cannot be confirmed(51).
The timing of treatment commencement has been pointed out as a plausible explanation for the failure(49). Due to the lengthy pre-clinical stage, drug administration after appearance of dementia may not be able to revert the deleterious impact of accumulated ?-amyloid in time(49).
Figure 4. Forest Plots generated from collated data from four trials (51)
Plots illustrate the mean difference in the (a) ADAS-Cog and (b)DAD scales, with overall effect depicted by the diamond at the end of each group. Results are insignificant if diamond crosses the line of no effect.
ADAS-Cog is utilized as the primary outcome in most trials but this scoring is reported to have poor sensitivity(52). It is not very accurate for measuring the decline in cognition and the weighted score can mask the positive effects of particular abilities(52). ADAS-CogIRT is shown to be an improved version that could be used in future trials(52).
Failure in all the previous ?-amyloid antibodies trials made researchers questioned the plausibility of the amyloid hypothesis. Interestingly, propitious results were obtained in the recent PRIME trial that led to the initiation of many trials aiming to further investigate its efficacy in AD(53).
The double blinded phase 1b (n=165) studied the effect of ADU versus placebo on safety and amount of cerebral ?-amyloid plaques in patients with milder AD(53). Subjects underwent a triple screening to confirm AD diagnoses and the trial excluded individuals with co-morbidities that could confound evaluation. Decline in the plaques, which correlated with increasing doses, was evident from the PET scans(53). As interpretation of the scans was subjective, using results from two qualified neurobiologists attempted to minimize human error. The CDR-SB and MMSE scales assessed cognitive deterioration(53). Both exhibited a decreasing dose-dependent trend on AD worsening at week 54 though it was only significant at the highest concentration (10mg/kg) for CDR-SB and at 3 and 10mg/kg (Figure 5)(53). However, the trial was not adequately powered so influences on cognitive performances and AD progression were
Figure 5. Changes in cognitive scores when treated with Aducanumab(53)
Bar Charts show the effects of Aducanumab on (a)CDR-SB and (b)MMSE scores at week 24 and 52. * denotes P